an epitope common to gangliosides 0-acetyl - cancer research

[CANCER RESEARCH 49, 3891-3897, July 15, 1989]

An Epitope Common to Gangliosides 0-Acetyl-GD3 and GD3Recognized byAntibodies in Melanoma Patients after Active Specific Immunotherapy1

Mepur H. Ravindranath,2 Donald L. Morton, and Reiko F. Irie

Division of Surgical Oncology, John Wayne Cancer Clinic, and Armand Hammer Research Laboratories, Louis Factor Bldg 9-267, University of California School ofMedicine, Los Angeles, California 90024

ABSTRACT

GDIis a major ganglioside of human melanoma and was shown to bean effective target for passive immunotherapy with murine monoclonalantibodies. It was noted earlier that G,,.,neither purified nor in melanomacell vaccine (MC V), could elicit an antibody response in melanomapatients. In this study, we demonstrate that melanoma patients whoreceived MCV had autoantibodies against a derivative of '.â€¢â€ž>,0-acety-lated GDI(6>-Ac(;,,,), a minor ganglioside expressed on human melanomacells, and that the antibodies cross-reacted with Gnl. Thin layer Chromatographie immunostaining revealed that all of the sera containingantibodies against O-AcGoj also reacted to G,,,. None of the other seraresponded only to Gm, although the MCV contained 7- to 12-fold higherGDJ than 0-AcGM. Furthermore, the antibody activity was completelyabolished by absorption with animal erythrocytes expressing either ()-acetyl disialogangliosides or GDI, indicating that the antibodies recognizean epitope commonly shared by (.,,, and O-\cGm. The antibodies boundonly to the sialyloligosaccharide moiety but not to the ceramide portionof GDI after endoglycosylceramidase treatment. The antibodies failed tobind to Gi)t after neuraminidase treatment. These results indicate thatthe sialyloligosaccharides of the gangliosides are important componentsof the epitope. Periodate oxidation abolished reactivity of the antibodiesto GD.Ibut not that to 0-AcGDj, revealing that the glycerol side chain ofthe sialic acids in both Gi)ts was an important structure of the epitope.The binding of the antibodies to melanoma cell surface gangliosides wasconfirmed by an absorption with a GDI-and 0-AcGDJ-positive melanomacell line. These results in the light of previous reports on the inability ofGDIto elicit immune response in humans suggest that Â¡iiiti-CDIantibodiesfound in the melanoma patients were induced by immunization with O-AcGoj and 0-AcGD.i present in the MCV would serve as an antigensource for Goj-targeted active specific immunotherapy of melanoma.

INTRODUCTION

Gangliosides are recognized as potential target antigens inimmunotherapy of melanoma (1). These glycolipids contain ahydrophobie ceramide portion embedded in the cell membraneand a hydrophilic sialyloligosaccharide moiety exposed (1).Only a small but sequential group of gangliosides is expressedon the melanoma cell surface (1). Normal melanocytes, precursors of melanoma, predominantly contain GM,3 (>95%), a

minor fraction of GD.i (<5%), and trace amounts of GM2 (2).On neoplastic transformation, the GM,:GD:<ratio changes from19:1 up to 1:16 (2-4). Concomitant with the increase in GDj,two other gangliosides appear on melanoma cell surface. Theyare GD2 (2) and O-AcGm (5). O-AcGm has not yet been biochemically demonstrated in any normal human tissues.

Since GD.I(see structure in Fig. 1) is a major ganglioside that

Received 11/14/88; revised 3/27/89; accepted 4/4/89.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This work is supported by USPHS Grants CA-12582, CA-36064, and CA-

42396 from the National Cancer Institute.2 Recipient of the Cancer Research Institute Fellowship Award.3The abbreviations used are: GMj. Siaa2-3GalÃÂ¡l-4Glc-ceramide; GDI, Siaa2-

8Siaa2-3Gal01-4Glc-ceramide; GM2, SiaÂ«2-3(GalNAc/31-4)Gal/31-4GIc-cer-amide; GD2, SiaÂ«2-8Sian2-3(GalNAcOl-4)Gal/Jl-4Gal-ceramide; O-AcGD3, O-acetylated GD3; MCV, melanoma cell vaccine; ELISA, enzyme-linked immuno-sorbent assay; TLC, thin layer chromatogram; IA, ummunoadherence; PBS,phosphate-buffered saline.

is prominently expressed on melanoma cell surface, it was usedas a target for passive immunotherapy of melanoma using GD1-binding murine monoclonal antibodies (7, 8). Melanoma patients were also immunized with purified GD3 to induce activeimmunity in patients, but no detectable anti-GD.i antibodieswere observed (9). Early attempts at immunization trials utilizing a MCV containing all of the melanoma-associated gangliosides, elicited antibody response against Gn2 and GM: but notagainst GD.i(10), indicating that GnÂ¡may not be immunogenicin humans. However, autoantibodies against 0-AcGi>,, an O-acetylated derivative of GD.,(see structure in Fig. 1), was notedin the sera of immunized melanoma patients (5); a unique lectinthat specifically binds to O-acetylsialic acids on gangliosidesblocked the binding of the antibodies from the sera of thepatients to a melanoma cell line (UCLASO-M25) that expresses prominently 6>-AcGD.,(5). In addition, sera of some ofthe MCV recipients contained antibodies reacting to purifiedO-AcGD1 in an ELISA.

While demonstrating the direct binding of such antibodies toO-AcGm on a TLC, we observed that those sera containingantibodies reacting to 0-AcGm also cross-reacted with GD.t.O-AcGD, is only a minor fraction of melanoma (rarely exceeding10% of total gangliosides), whereas Gni ranges between 20 and80% of the total gangliosides of melanoma (2-4). Therefore,GDJ would be an excellent target for the active specific immunotherapy of melanoma, if the sources of the antigen responsible for the induction of GD.i antibodies in patients can beidentified. Here we report the antibody response against GD.,and have assessed the possibility of the cross-reactivity of anti-O-AcGD, antibodies with GD1. This study postulates that O-acetylation may render Gâ„¢ immunogenic in melanoma patients.

MATERIALS AND METHODS

Antibody Source. The sera of melanoma patients collected before andafter immunization were used as the source of antibody. Sera werestored at â€”¿�80Â°Cand then thawed and centrifugea before being used atIO4x g at 4Â°C.We selected 20 of the melanoma patients based on the

positivity of their sera to UCLA melanoma cell line M25 in IA (11)and membrane immunofluorescence (12) assays. Fourteen of thesepatients received MCV consisting of three different melanoma celllines, after surgery. In addition we have used the sera of three patientsthat were negative to IA against M25 as a control.

Two kinds of MCVs were given to the patients intradermally ( 10) orintralymphatically (13) during the course of the clinical trials. Prior touse, the MCV was irradiated with IO4rad in a 2 s-a dose/time schedulethat maintains cell viability but prevents cell proliferation. For intralym-phatic immunization a total of 25 x 10" cells in 5 ml of lactatedRinger's solution were distributed equally at one to three sites; generally

two sites were used for each immunization. The dorsal aspect of eitherthe hands or the feet was the site of lymphatic cannulation. Forintradermal immunization a total of 75 x IO6cells were administered

at multiple sites on the lower abdomen and lateral chest. Table 1 showsthe UCLA melanoma cell lines used in the vaccine and the relativeratio of the gangliosides in the vaccine.

Ganglioside: Source, Purification, and Analysis. The gangliosides of3891

on April 9, 2019. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

ANTIBODY RECOGNITION OF O-AcGD) AND GD, IN MELANOMA PATIENTS

M25 melanoma cell line were isolated and purified following theprocedure of Ladisch and Gillard (14) and modified by us (5). Thegangliosidcs purified by gel filtration were subjected to TLC. TLC wasperformed with chloroform:methanol:0.25% Cad: in water (60:30:9,v/v/v). The glycolipids were detected with orcinol-ferric chloride andresorcinol-HCI. The other gangliosides used in this study includedbovine brain G,,, (supplied by Dr. Pallman, K. G. 800 Munich 40, WestGermany) and bovine brain GM1(Sigma). The gangliosides are identified by the nomenclature of Svennerholm (15).

Enzymatic Alteration of Gangliosides. The oligosaccharide moiety ofCDÂ«was enzymatically cleaved from the ceramide portion using a novelenzyme endoglycosylceramidase (GeneZyme, Boston, MA), purifiedfrom Rhodococcus (16). Fifty Mgof dÂ» and 0.5 milliunit of endoglycosylceramidase were incubated in 50 HIMsodium acetate buffer (pH6.0) containing 0.5 mg/ml of sodium taurodeoxycholate (total volume,50 n\) for 72 h at 37*C. After the incubation, the hydrolysate was

diluted with 9 volumes of methanol, evaporated to dryness over N2,and resuspended in chloroforrmmethanol (1:1, v/v).

The sialic acids of G()1were removed using Clostridium perfringensneuraminidase in the presence of sodium taurocholate (5). We have notused O-AcGi), for neuraminidase treatment because it is known that O-acetylation renders the sialic acid resistant to C. perfringens neuraminidase. Fifty jig of GDI and 150 /amits of neuraminidase wereincubated in 50 mM sodium acetate buffer (pH 5.5) containing 0.5 mg/ml of sodium taurodeoxycholate (total volume, 50 n\). After incubation,the hydrolysate was diluted with 9 volumes of methanol, evaporated todryness over N:, and resuspended in chlorofornrmethanol (1:1, v/v).

Base Treatment of Gangliosides. O-Acetyl groups of O-AcGi>, wereremoved by base treatment with concentrated ammonium hydroxide (1h at 25Â°C)as described earlier (5). Ammonia was evaporated over N;;

the gangliosides were washed three times with methanol and evaporatedover N2.

Periodic Acid Oxidation. To cleave the glycols in the glycerol sidechain of G,>,s, the gangliosides were incubated in 5 HIMsodium meta-periodate in PBS for l h at 37Â°C.The periodate-oxidized gangliosides

were then chromatographed and tested for antibody binding.IA. The IA, performed as described by Irie et al. (Il ), is a sensitive

method for detecting antibody bound to human tumor cells. In thisassay target cells were coated with antibodies by incubating the cellswith varying dilutions of antibodies. After 90 min of incubation at37'C, the cells were washed and incubated with a guinea pig fresh

Ceromide

Sio2

Fig. 1. Slructures of melanoma-associated Cn>,and O-AcGnl. Note that O-acetylation is at C-9 of ihe terminal sialic acid (Sia 2) in O-AcGDi [based on thework of Thurin et al. (6)].

Table Gangliosiiie profiles of MCV I (Ml 4. M 12, M7) and II (MI4. M20,M7)

GangliosidesGMJGD)GM2GD2O-AcGD,Â°%of

totalgangliosidesMCV

I MCVII4036II)103402027103

" Identified by R,. alkali lability, and comparison with the Rf of O-AcGD\

purified and isolated from M25 cells.

3892

serum (complement) (diluted 1:40) for 10 min. After complete settlingof the mixture of melanoma cells and erythrocytes (20 min), erythrocyteresetting around target cells were examined. The percentage of rosette-forming tumor cells is proportional to the antibody titer in the sera. Inthis study UCLASO-M25 melanoma cells were used as the target cellsbecause of the cell surface expression of 0-AcGD.i by the cell line (5).

Immunofluorescence. Immunofluorescence was performed accordingto standard procedures described earlier (12). Fluorescein isothio-cyanate-coupled mouse anti-human IgG was used to visualize theantibody-binding sites. At least 1000 cells were analyzed in each experiment.

TLC Immunostaining and Autoradiography. The antibody binding toO-AcGi>i and Gr>, were identified on TLC using radioiodinated goatanti-human IgG antibodies (17, 18). The gangliosides (<5 Â¿ig)purifiedfrom M25 melanoma cells, bovine brain GD3 (2 ^g), or chemicallyaltered Gn, were chromatographed on alumina-backed high-performance TLC plates (Silica Gel 60; E. Merck). The chromatograms weredried and coated with 0.4% polyisobutylmethacrylate (l g dissolved in10 ml of chloroform and made up to 250 ml) in n-hexane for 30 s andtreated with 1% human serum albumin in PBS (pH 7.0) at roomtemperature for 2 h. The plates were then overlaid with diluted humansera (1:200 to 1:400) overnight at 4Â°C.After repeated washing in 1%

human serum albumin-PBS, the plates were overlaid with radioiodinated (125I)goat anti-human IgG for 4 h at 4Â°C.After repeated washing,

the plates were exposed to XR-5 x-ray film (Eastman Kodak, Rochester,NY) against an intensifying screen for at least 4 days at room temperature.

Enzyme-linked Immunosorbent Assay (ELISA). We performed theELISA for anti-ganglioside antibodies to detect ;mii (.,,> antibodies inthe patient's sera, following a method empirically assessed in our

laboratory for the monitoring of anti-ganglioside antibodies in cancerpatients (19). Assay wells were coated with 4 ^g of bovine brain GD.ÃŽ.The gangliosides were diluted to the desired concentration in absoluteethanol. Aliquots (50 Â¿J)were pipeted into each microtiter well andevaporated to dryness in a vacuum. Peroxidase-conjugated goat anti-human IgG or IgM was used as second antibody.

Absorption Assays to Assess Antibody Specificity. Absorption assayswere carried out to assess the specificity of serum antibodies to Cmand O-AcGnÂ«.The absorption of the antibodies by melanoma cells(M25) was initiated by incubating dilute sera (1:10) with an equalvolume ( 100 M')of packed live melanoma cells at 4Â°Cfor 18 h, followed

by centrifugation at 200 x g for 15 min. The absorbed sera (supernatant)were further diluted prior to analysis. The patients' sera (diluted at

1:10) were also absorbed to a 5-fold volume of packed fresh erythrocytesof different species expressing different kinds of glycolipids.

RESULTS

Detection of Anti-CMcGÂ», Antibodies in Patients' Sera. Gan

gliosides extracted from the M25 human melanoma cell linewere used for the TLC radioimmunostaining to detect anti-O-AcGixt antibodies in the patients' sera. Earlier we have used an

O-acetyl sialic acid-specific lectin to show that M25 humanmelanoma expresses a higher level of O-AcG,,, than any othercultured melanoma cell line tested (5). Fig. 2 shows that themajor gangliosides of M25 melanoma cells are GM, and GDI,constituting 21 and 60% of the total gangliosides, respectively,and O-AcGi>, constitutes about 7%, whereas GMJ and GD: are3 and <6% of the total gangliosides, respectively. The 0-AcG,>,migrated as a doublet almost in the same position of GM2(<2%). The O-AcGixi isolated from the total gangliosides, migrated to a position identical to GI>, on TLC after base treatment, which removes O-acetyl groups from G,â€ž(for figure seeRef. 5).

A total of 23 serum specimens obtained from melanomapatients (see "Materials and Methods") were diluted at 1:400

or 1:250 and tested by the TLC immunostaining. Most of thesera reacted to a doublet migrating at a position corresponding



ANTIBODY RECOGNITION OF O-AcGD) AND GDJ IN MELANOMA PATIENTS

GD2GT,

21% <2% 7% 60% <1%<6%<3%

Fig. 2. Densitometric profile of the total gangliosides purified from a melanoma cell line (M25). Note the relative position and intensity of GDJ and 0-AcGD3. All bands are resorcinol positive. GMJ is the fast-moving glycolipid onTLC. 0-AcGoj is the alkali-labile ganglioside. Isolated 0-AcGDJ after basetreatment migrates identically to GD3(5).

Point ofâ€¢¿�*-opplicotion

D-1-VA D-8-SE B-1-HA

Fig. 3. Autoradiography showing Â¡mmunoreactivity of antibodies with GDjand 0-AcGoj in the sera of three melanoma patients before (Lane A) and after(Lane B) immunization with MCV. The sera were tested at a dilution of 1:250.*, Corresponds to position of GMJ.

to GMJ. The immunological and epitope specificity of the GMJreactivity of the antibodies is under investigation in our laboratory. However, similar antibodies have been identified previously as human erythrocyte autoantibodies (20). Eight of the23 sera reacted to O-AcGDi doublets. The same sera that reactedto 0-AcGoi also reacted to GD,. Interestingly none of the serawas positive to GDJ alone.

Differences in Reactivity between Preimmune and ImmuneSera. To examine whether the immunization of patients withMCV elicited antibody response against O-AcGm and GD1, wecompared the reactivity of the sera before and after the administration of the MCV. The MCV contained both GD, (20 to36%) and O-AcGâ€ž,(3%) (Table 1). The GD, content of MCVis lower than the level of GDj seen in most of the melanomabiopsies and cultured cell lines. Three patients whose immunized sera showed a high reactivity on to GDJ and 0-AcGDj onTLC were examined. Both pre- and postimmunized sera weretested at a dilution of 1:250 and the results are presented inFig. 3. None of the preimmune sera reacted to OAcGDj orGDJ, in contrast to the positive reactivity in the sera obtainedfrom immunized sera. Therefore, the GDj/0-AcGDi positivitymight not be due to cross-reactivity of human erythrocyte anti-

ganglioside antibodies known to occur in some normal individuals (20). As noted earlier, all sera reacted strongly with thefractions corresponding to GM3. Immune sera reacted moreintensely than the preimmune sera to this fraction. All theimmune sera showed distinct reactivity to 0-AcGoj and GDI(Fig. 3). The sera of Patient D-1-VA reacted to Cm and GTi aswell.

Demonstration of Specificity of the Antibodies for 0-AcGnjand GD.I.In order to assess the specificity of the antibodies toGo., and O-AcGo.i, the positive sera were tested against theM25 melanoma gangliosides after absorption with erythrocytescontaining these gangliosides. We selected three different kindsof erythrocytes; their major ganglioside profiles are defined inTable 2. Ox erythrocytes contained only those gangliosides withthe terminal Siaa2-3Gal sequence characteristic of GNU (21),whereas the pig erythrocytes contained both GM, and GDj in aratio of 1:1 (22). About 50% of the ox and pig erythrocytegangliosides contained Â¿V-glycolylneuraminicacid as terminalsialic acid (21, 22). Neither of these erythrocytes contained O-acetylsialic acids. Rat erythrocytes, on the other hand, contained O-acetylsialic acids in a disialoganglioside (23). Raterythrocytes do not contain any detectable quantity of GDI, andnone of the erythrocytes tested contains GD2(Table 2).

Fig. 4 presents the results of antisera reactivity on TLC beforeand after preabsorption. While the sera preabsorbed by oxerythrocytes continued to react with the doublets correspondingto GM.,, GDJ, O-AcGDj, and GD?, the sera preabsorbed to pigand rat erythrocytes failed to react with both GI>, and O-AcGDj.These results confirmed that immunostaining to areas of GDJand 0-acetylated gangliosides on TLC was in fact due to theantibody reactivity against these ganglioside antigens. Pig erythrocytes that contained GDJ but not any 0-acetylgangliosides,and the rat erythrocytes that contained 0-acetyldisialoganglio-side but not GDj abolished the reactivity of the sera to both 0-AcGoj and GDJ (Fig. 4), indicating the cross-reactivity of theantibodies reacting to both the gangliosides. The antibodies

Table 2 Major gangliosides of the erythrocytes used for absorption (based onKefs. 21 to 23)

Erythrocytes used for absorption

Gangliosides Ox" Pig Rat

0-AcGD" All the gangliosides contain terminal sequence characteristic of (,,.,,. with

acetylneuraminic acid and glycolylneuraminic acid in a ratio of 1:1. +, relativeabundance of the gangliosides on the erythrocytes; â€”¿�.their absence.

O-AcGDj-Q

Pointofâ€¢¿�*-application

Fig. 4. Autoradiography showing immunoreactivity of antibodies with GDIand 0-AcGoj in the serum of an immunized patient (Lane B) compared with apreimmune serum (Lane A), the immunized serum adsorbed with Gnj-negativebut Gun-positive (ox; Lane Cl), GD,-rich (pig; Lane 2). and Cm-negative but 0-acetyldisialoganglioside-rich (rat; Lane C3) erythrocytes. *. corresponds to posi

tion of GMJ.

3893



ANTIBODY RECOGNITION OF O-AcGDJ AND GDJ IN MELANOMA PATIENTS

reacting to G,,: were unaffected by erythrocyte absorption treatments, suggesting that the Gni-reacting antibodies are distinctfrom the anti-GI)2 antibodies. Erythrocyte-absorbed serashowed a decrease in reactivity to the doublet corresponding toGNU,which requires further investigation with regard to antibody specificity.

Epitope Analysis of <;,,i- and O-AcGiÂ»-Reactive Antibodies.Gangliosides consist of hydrophilic (carbohydrate portion) andhydrophobic (ceramide, including the sphingosine and fatty acidportions) regions (Fig. 1). The ceramide portion is embeddedinto the bilayered lipid membrane of the cell, whereas thecarbohydrate portion is exposed outside the membrane. Antibodies binding to the ceramide portion have no immunothera-peutic potential, inasmuch as the ceramide portion is not exposed on tumor cells. To exclude the possibility that the antibody cross-reactivity between GD, and O-AcGn.i is due to acommon antigen on the ceramide portion of both gangliosides,the oligosaccharide moiety and ceramide of G,)( were testedindependently by cleaving the linkage between them with anovel enzyme called endoglycosylceramidase (16). On TLCplates, the ceramide (stained brown in orcinol-FeCl2), migratedfaster than GD, and the oligosaccharide moiety (NeuAc<*2-8NeuAcÂ«2-3Gal01-4Glc-) (stained purple in orcinol-FeCb),remained just above the origin in the solvent system used. Theantibodies reacting to Gn< bound only to the oligosaccharideportion of the gangliosides but not to the ceramide portion,suggesting that the epitope of the antibodies is located on theoligosaccharide moiety of GD,. The antibody binding may require all the sugar residues in the carbohydrate portion or onlythe terminal of the oligosaccharide. We tested the reactivity ofthe antibodies by cleaving the disialyl residues of the terminalportion of the oligosaccharide with C. perfringens neuramini-dase (Sigma type X). The antibodies did not react to the asialo-GD.I, indicating the specific affinity of the antibodies to thedisialic acid residues characteristic of Gm and O-AcGD3.

We have further characterized the epitope determinants ofthe antibodies in the following experiments. We have subjectedM25 gangliosides to two kinds of chemical treatments, namelybase treatment and periodate oxidation. We have previouslyshown base treatment of 0-AcGn> removes the O-acetyl groupsfrom the sialic acids constituting the gangliosides (5). If alkalitreatment eliminates the TLC immunostaining of the 0-AcGn,band, then it is evident that the antibodies are directed againstthe O-AcGin structure. The results presented in Fig. 5 clearlydemonstrate that alkali treatment eliminated the TLC immunostaining of the 6>-AcGi,i band indicating that the antibodies arein fact directed against O-AcGm.

Periodate treatment gives additional proof of the antibodyspecificity for both O-AcG,,, and GD,. If the glycerol side chainof sialic acids are involved in antibody binding then cleavingthe glycols in the glycerol side chain should abolish immunostaining of GI><;however, periodate oxidation will not abolishthe immunostaining of 0-AcGm because the substitution of O-acetyl groups at C-9 would prevent oxidation of glycols. Theresults obtained with periodate oxidation is more interesting inthat none of the gangliosides except O-AcG,Â»was recognizedby the antibody, demonstrating the importance of the glycerolside of the terminal sialic acid of GDI and O-AcGm in antibodyrecognition.

Demonstration of the Binding of Anti-Gi>, or Anti-0-AcGi>,Antibodies to the Cell Surface of Melanoma Cells. The bindingability of the patients' serum anti-dÂ» or 0-AcGn, antibodies

to melanoma cell surface was assessed by absorption of theantibodies with M25 melanoma cells that contained GD, and

GM,-

'J

OO

CU o>

B o1 o

O> o>

m

"Z oo>

Q_

Fig. 5. TLC immunostaining of M25 gangliosides with the sera of melanomapatient D-I-VA. after base treatment and periodate oxidation. Note that basetreatment eliminates immunostaining of the O \Â¡-(,,,.band and periodate treatment eliminates the <-,.. band but not the f' V(,,,, band. The 0-acetyl group onthe glyeerol side chain of sialic acids of O-AcGD, is base labile. O-Acetylation onthe glycerol side chain renders glycols resistance to periodate oxidation.

0-AcGi>.i in a ratio of 9:1. Reduction of antibody reactivity topurified GDI would be indicative of the binding of the antibodiesto the cell surface antigens. Sera were tested for both IgM andIgG antibodies using ELISA. Initially we have tested the specificity of the ELISA absorption assay using erythrocytes knownto express Gm or devoid of GD., on their cell surface, namelypig and ox erythrocytes, respectively. Sera were tested for bothIgG and IgM antibodies after absorption with erythrocytes. Asshown in Fig. 6, sera preabsorbed with ox erythrocytes showedan undiminished dosimetrie response to GDI with the ELISAat varying dilutions. The reactivity of the sera preabsorbed withpig erythrocytes was remarkably reduced even at high concentrations of the sera. This preliminary test indicates the validityof the assay for demonstration of anti-Cm antibody binding tocell surface GD,. In the M25 melanoma cell absorption, all ofthe sera were preabsorbed initially with ox erythrocytes. Thisis particularly important because melanoma cell lines grown ina medium containing fetal bovine serum can express gangliosides containing Â¿V-glycolylsialicacids (24). In order to avoidcross-reactivity of the antibodies with jV-glycolylsialic acid containing gangliosides on melanoma cell surface, the serum waspreabsorbed to ox erythrocytes which predominantly expressgangliosides with A'-glycolylsialic acids (21).

As shown in Table 3, both IgM and IgG antibodies showedremarkable reduction in their reactivity to GD, after absorptionto M25 melanoma cell, confirming the ability of the antibodiesto bind to the tumor cells directly.

DISCUSSION

In this investigation we tested sera from melanoma patientswho received MCV immunotherapy for the presence of anti-0-

3894



ANTIBODY RECOGNITION OF O-AcCDJAND GD1 IN MELANOMA PATIENTS

B-2-HO: anti-GD3-IgM

ELISA ODat 492 ni

I: NO I: 1600 1:Â£00

DilutionofAntisera

ELISA00 .at492nu

i.i.

0-8-SE:anti-G03-IgG

1 '00 I 900 1: IMO 1:3200 I 6400

Dilution of Antisera

Fig. 6. Reactivity of antibodies to purified Gm in ELISA, after absorption ofpatient sera with Goi-negative ox erythrocytes (A) and GD,-rich pig crythrocytes(B).

AcGn., and ;mti <â€¢,,;antibodies and observed that antibodiesreacting to G,,, occurred only in the sera containing anti-O-AcGn, antibodies. MCV contained 7- to 12-fold higher G0.,than O-AcGn.,. These findings led to the inference that GD1

might not have elicited an immune response. This suggestiongains support from the works of Livingston (9) who was notable to induce anti-Gni antibodies in melanoma patients usinga purified GDI vaccine. The unresponsiveness against GI>, doesnot appear to be due to the purified form of the immunogenbecause the same purified GDI vaccine was found to be immu-

nogenic when used to immunize mice. The same group has alsoobserved that purified GM2vaccine was immunogenic in melanoma patients (25). It is also highly unlikely that the immunization methodology was inappropriate, in the light of theseinvestigators' previous success in inducing anti-GM: and anti-

Go: antibodies in melanoma patients using the same administration techniques (25). Another indication that GD, might nothave elicited the observed immune response is its occurrence inlarge quantity in normal tissues. GÂ».iaccounts for over 20% ofthe total gangliosides in several normal tissues as shown inTable 4. Table 4 indicates that Gn., is a normal component inhumans and thus humans may have acquired immunologicaltolerance against GDI. Finally, if GDI were immunogenic inhumans, a melanoma tumor would not have escaped fromcomplement-mediated antibody cytotoxicity and antibody-dependent cellular cytotoxicity (36) because GH.,is so prominentlyexpressed on the melanoma cell surface.

The evidence that the anti-GÂ».,antibodies found in MCV-treated patients may have been induced by the antigenic stimulation of O-AcGi>., contained in the MCV is: (a) antibodieswere not found in the preimmune sera of immunized patients;(b) there is a paucity of evidence demonstrating the presence of0-AcGu., in normal human tissues; (c) previous reports demonstrate the nonimmunogenic nature of GD, in humans; (d)there are no anti-GÂ».,antibodies alone in any of the immunizedsera analyzed in spite of the higher amount of GDI than O-AcGi>.,;and (e) the anti-GÂ».,antibody activity was abolished byabsorption not only with GD<but also with O-acetylsialic acid-containing cells suggesting that the antibody detects a commonepitope between GDI and O-AcGÂ».,.Moreover a recent study byHirabayashi et al. (24) indicates that chemical alterations in thesialic acid residues of gangliosides can render the moleculeimmunogenic. It was demonstrated in this study that the gangliosides substituted with the jV-glycolyl group in the positionof yV-acetyl group elicited an antibody response against thosegangliosides, which are normally nonimmunogenic (24). If sucha modification can render immunogenicity to a nonimmunogenic molecule on the cell surface, then such alterations couldbe of immense value in the immunotherapy of cancer, particularly for those tumors that express large quantities of cellsurface molecules which by themselves are not foreign to thehost.

The immunostaining results of GDI degraded by endoglyco-sylceramidase and neuraminidase have provided evidence thatthe sialic acid residues of GDI and 0-AcGi>., are the importantcomponents of the epitope recognized by these antibodies.Although the disialyl residues occur in GD:, the erythrocyteabsorption studies reveal that the antibodies recognizing GD:are not absorbed by erythrocytes expressing GDI and hence theepitope requirement of GDI and 0-AcGn., antibodies are differ-

Table 3 ELISA immunostaining of antibodies reading lo dâ€žhe/ore and after absorption with .1/25 human melanoma cells

For this assay the wells were coated with 4 ^g of bovine brain GD, diluted in absolute ethanol. Aliquots (50 Â¿il)were added to each well and evaporated to drynessin vacuum. Peroxidase-conjugated anti-human IgG or IgM were used as second antibodies. Sera were preabsorbed to I ml of packed ox erythrocytes. For absorptionwith M25 cells, sera were initially preabsorbed with 0.1 ml of packed ox erylhrocytes followed by 0.2 ml of packed melanoma cells, melanoma cells were removedfrom culture flasks after trypsinization. Values are expressed as means of triplicate analyses.

SeraA-8-BRB

2-HOD8-SETreatments

(prcabsorption)Ox

RBCM25cellsOx

RBCM25cellsOx

RBCM25 cellsIgM1:8000.940

0.4700.700

0.4201:16000.850

0.2300.580

0.1601:4000.700

0.0851.160

0.640IgG1:8000.620

0.0801.100

0.6401:16000.415

0.070

3895



ANTIBODY RECOGNITION OF O-AcGD3 AND GDJ IN MELANOMA PATIENTS

Table 4 Normal human tissues showing high level o/CDJ

Nature of tissuesMilk"RetinaDura

materThymusFetal2

moDermisSubcutaneousThyroidAorta

intimaColoniemucosaKidney%

of totalgangliosides53-74SO50322231242625232019Refs.262728293030313132333435

" On days of lactation 2 to 40.

ent from those which bind to G,,:. Antibodies failed to bind toGDJ after periodate oxidation but reacted to periodate-treated0-AcGnj, which is resistant to periodate oxidation. This findingdemonstrates that the antibodies recognize both substituted(with an 0-acetyl group) and unsubstituted glycerol side chainof sialic acids on the ( i,,,s and do not bind to the gangliosidesif the side chain is destroyed by periodate oxidation.

One question that arises from this investigation is the consequences of inducing an immune response to a nonimmuno-genic molecule commonly found on normal cells. This is aninteresting point, especially in light of the clinical trial with amurine monoclonal anti-G0j antibody (R-24) which did notdemonstrate direct cytotoxic effects to normal tissues containing GDJ. Why normal cells expressing GDJ are not affected bythis antibody is still unknown, but it has been hypothesized thatthe antigens on the nonreactive cell surface are immunologicallycryptic. If this is the case, O-AcGDj might be an effective meansfor inducing cross-reactive antibody since it is found in amajority of biopsied melanomas (83%, n = 52) (4). Since O-AcGÂ»,is a minor component on melanoma cells (<5% of thetotal gangliosides), it may be insufficient to induce a strongimmunity in melanoma patients without a deliberate immunization. However, if 0-AcGnj can induce antibodies that reactto melanoma-associated GDJ, the immunization of patients withO-AcG[>j may be a useful approach to manipulating melanomacell growth in vivo. In addition, the accumulation of GDj hasbeen reported for other human malignancies, such as astrocy-toma (37), glioma (38), meningioma (28, 29), neurofibrosar-coma (31), and acute nonlymphoblastic (39) and chronic lym-phocytic leukemia (40). Therefore the model developed for themelanoma system may apply to other human cancers.

ACKNOWLEDGMENTS

We thank Peter Chandler for his valuable suggestions.

REFERENCES

1. Ravindranath, M. II.. and Irie, R. F. Gangliosides as antigens of humanmelanoma. In: I . Nathanson (ed.). Malignant Melanoma: Biology, Diagnosisand Therapy, pp. 17-43. Boston: Kluwer Academic Publishers, 1988.

2. Carubia, J. M., Vu. R. K.. MÃ¡cala.L. }., Kirkwood, J. M., and Varga, J. M.Gangliosides of normal and neoplastic human melanocytes. Biochem. Iliophys. Res. Commun.. 120: 500-504. 1984.

3. Herlyn. M.. Thurin, J., Balaban. G., Bennicelli, J. L., Herlyn, D., Eider, D.E., Bondi. E., Guerry, D., Nowell. P., Clark, W. H., and Koprowski, H.Characteristics of cultured human melanocytes isolated from different stagesof tumor progression. Cancer Res.. 45: 5670-5676. 1985.

4. Tsuchida, T., Saxton, R. E., Morton, D. L.. and Irie. R. F. Gangliosides ofhuman melanoma. J. Nati. Cancer Inst., 78: 45-54, 1987.

5. Ravindranath, M. H.. Paulson. J. C.. and Irie. R. F. Human melanomaantigen 0-acetylated ganglioside <â€¢¿�,,<is recognized by Cancer anlennariuslectin. J. Biol. Chem., 263: 2079-2086. 1988.

6. Thurin. J., Herlyn, M.. Hindsgal, O.. StrÃ¶mberg, N. Karlsson, K-A., Elder,

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

D., Steplewski, Z., and Koprowski. H. Proton NMR and fast-atom bombardment mass spectroscopy analysis of the melanoma-associated ganglioside9-O-acetyl-Gn,. J. Biol. Chem., 260: 14556-14563, 1985.Houghton, A. N., Mintzer, D., Cordon-Cardo, C, Welt, S., and Fliegel. B..Vadhan. S., Carswell. E., Melamed, M. R., Oettgen, H. F., and Old, L. J.Mouse monoclonal IgG3 antibody detecting GDJ ganglioside: a phase I trialin patients with malignant melanoma. Proc. Nati. Acad. Sci. USA, 82:1242-1246, 1985.Goodman, G. E., HellstrÃ¶m, I., Hummel, D., Brodzinsky, L., Yeh, M. Y.,and HellstrÃ¶m. K. E. Phase I trial of monoclonal antibody MG-21 directedagainst a melanoma associated GDj ganglioside antigen. Proc. Am. Soc. Clin.Oncol., 6: A823. 1987.Livingston, P. O. Immunization with purified gangliosides. J. Cell Biochem..Â¡2E:123. 1988.Tai, T., CaÃ±an,L. D., Tsuchida, T., Morton, D. L., and Irie, R. F. Immu-nogenicity of melanoma associated gangliosides in cancer patients. Int. J.Cancer, 35:607-612, 1985.Irie, R. F., Irie, K.. and Morton, D. L. Natural antibody in human sera to aneoantigen in human cultured cells grown in fetal bovine serum. J. Nati.Cancer Inst., 52: 1051-1058, 1974.Irie. K., Irie, R. F., and Morton, D. L. Humoral immune response tomelanoma-associated membrane antigen and fetal brain antigen demonstrated by indirect membrane immunofluorescence. I. Cancer Immunol.Immunother., 6: 33-39. 1979.Ahn, S., Irie, R. F., Weisenburger, T. H.. Jones. P. C.. Juillard. G., Roe, D.J., and Morton, D. L. Humoral immune response to intralymphatic immu-notherapy for disseminated melanoma: correlation with clinical response.Surgery, 92: 362-367, 1982.Ladisch, S., and Gillard, B. A solvent partition method for microscaleganglioside purification. Anal. Biochem., 146: 220-231, 1985.Svennerholm, L. Chromatographie separation of human brain gangliosides.J. Neurochem., 10:613-623, 1963.Ito, M., and Mamagata, T. A novel glycosphingolipid-degrading enzymecleaves of the linkage between the oligosaccharide and ceramide of neutraland acidic glycosphingolipids. J. Biol. Chem., 261: 14278-14282, 1986.Hansson, G. C., Karlsson, K. A., Larson, G., Samuelsson. B. E., Thurin, J.,and Bjursten. L. M. Detection of blood group type glycosphingolipid antigenson thin-layer plates using polyclonal antisera. J. Immunol. Method, 83: 37-42. 1985.Saito, M., Kasai, N., and Yu, R. K. In situ immunological determination ofbasic carbohydrate structures of gangliosides on thin layer plates. Anal.Biochem., 148: 54-58, 1985.Tai, T., Cahan. L. D., Paulson, J. C., Saxton, R. E., and Irie, R. F. Humanmonoclonal antibody against ganglioside (.,,-: use in development of anti-Co; in cancer patients. J. Nati. Cancer Inst., 73:627-633, 1984.Uemura, K., Roelcke, D., Nagai, Y., and Feizi, T. The reactivities of humanerythrocyte autoantibodies anti-Pr 2, anti-GD, Fl and Sa with gangliosidein a chromatogram binding assay. Biochem. J., 219: 865-874, 1984.Chien, J., Li, S., Laine, R. A., and Li, Y. Characterization of gangliosidesfrom bovine erythrocyte membranes. J. Biol. Chem., 253:4031-4035. 1978.Suzuki, Y., Suzuki, N., Miehi, H., and Matsumoto, M. Characterization ofgangliosides of porcine erythrocyte membranes: occurrence of gangliosideGDJ as major ganglioside. Lipids, 20: 588-592. 1985.Gowda, D. C.. Reuter, G., Shukla. A. K., and Schauer, R. Identification of adisialo-ganglioside (Gm.,) containing terminal V-ace(vl-9-O-acetyl neura-minic acid in rat erythrocytes. Hoppe-Seyler's Z. Physiol. Chem., 365:1247-

1253, 1984.Hirabayashi, Y., Higashi, H., Kato, S., Taniguchi, M.. and Matsumoto, M.Occurrence of tumor-associated ganglioside antigen with Hanganutziu-Deicher antigenic activity on human melanomas. Jpn. J. Cancer Res., 78:614-620, 1987.Livingston, P. O., Natoli, E. J., Calves, M. J., Stocken, E., Oettgen, H. F.,and Old, L. J. Vaccines containing purified GM> ganglioside elicit ('<-.,â€¢

antibodies in melanoma patients. Proc. Nati. Acad. Sci. USA, 84: 2911-2915. 1987.Takamizawa, K., Iwamori, M., Mutai, M., and Nagai, Y. Selective changesin gangliosides of human milk during lactation: a molecular indicator for theperiod of lactation. Biochim. Biophys. Acta, Â«79:73-77, 1986.Holm, M., Mansson, J-E., Vanier, M-T., and Svennerholm, L. Gangliosidesof human, bovine and rabbit retina. Biochim. Biophys. Acta, 280: 356-364,1972.Berra, B., Riboni. L., De Gasperi, R.. Gaini, S. M., and Ragnotti, G.Modifications of ganglioside patterns in human meningiomas. J. Neurochem., 40: 777-782, 1983.Sunder-Plassman. M., and Bernheimer. H. Ganglioside in meningiomen andheinhauten. Acta Neuropathol. (Beri.). 27: 289-294, 1974.Etienne-Decerf. J.. Gosselin-Rey, C., Gosselin, L., and Winand. R. Effectsof thyrotropin and thyroid stimulating immunoglobulins on gangliosideslabeling of human thymus cultured pathological cells. J. Endocrino!. Invest.,9:57-63, 1986.Tsuchida, T., Otsuka, H.. Niimura. M., Inoue, M., Kukita, A., Hashimoto,Y., Seyama, Y., and Yamakawa, T. Biochemical study on gangliosides,neurofibromas and neurofibro-sarcomas of Recklinghausen's disease. J. Der-matol.. //: 129-136, 1984.Bouchon, B., Portoukalian, !.. and Bornet. H. Major gangliosides in normaland pathological human thyroids. Biochem. Int., 10: 531-538, 1985.Prokazova. N. V., Orekhov, A. N., Mukin, D. N., Mikhailenko, 1. A., Kogter,L. S., Sadovskaya. V. L., Golovanova, N. K., and Bergelson. L. D. The

3896



ANTIBODY RECOGNITION OF O-AcGD, AND GD) IN MELANOMA PATIENTS

gangliosides of adult human aorta: intima, media, and plaque. Eur. J. 37. Berra, B., Caini, S. M., and Riboni, L. Correlation between gangliosideBiochem., 767: 349-352, 1987. distribution and histolÃ³gica!grading of human astrocytomas. Int. J. Cancer,

34. Hakomori, B., Nudelman, E., Levery, S. B., and Patterson, C. M. Human 56'363-366 1985cancer-associated gangliosides defined by a monoclonal antibody (IB9) di- 38 Fredmani p.', von Ho|st< B , Collines, V. P., Ammar, A., Dellheden. B.,

reeled to s.alosyl a2-6galactosyl residue: a preliminary note. B.ochem. B,o- Wanren B Granholm. L and Svenncrholm, L. Potenlial gang|ioside anti_pnys. ixCS. i oiMiniiii., iij. f~fi â€”¿�/~to. ivoj. â€¢¿�. â€¢¿�. ,. .. ., , _ â€ž¿�,_- .^^ n*\f

35. Rauvala,H. Gangliosides of human kidney. J.Biol.Chem.,25/: 7517-7520, â€ž¿�gens associated w,th human gl.omas.Neurol. Res., A; 123-126, 19861976 39. Westnck, M. A., Lee, W. M. F., Goff, B., and Macher. B. A. Gangliosides

36. Cheresh, D. A., Honsik, C. J., Staffeleno, L. K., Jung, G., and Reisfeld, R. of human acute leukemia cells. Biochim. Biophys. Acta, 750:141-148, 1983.A. Disialoganglioside GDJ on human melanoma serves as a relevant target 4<>.Goff, B. A., Lee, W. M. F., Westrick, M. A., and Macher, B. A. Gangliosidesantigen for monoclonal antibody-mediated tumor cytolysis. Proc. Nati. Acad. of human chronic lymphocytic leukemia and hairy cells. Eur. J. Biochem.,Sci. USA, 82: 5155-5159, 1985. 130: 553-557, 1983.

3897



1989;49:3891-3897. Cancer Res Mepur H. Ravindranath, Donald L. Morton and Reiko F. Irie Specific ImmunotherapyRecognized by Antibodies in Melanoma Patients after Active

D3 and GD3-Acetyl-GOAn Epitope Common to Gangliosides

Updated version

http://cancerres.aacrjournals.org/content/49/14/3891

Access the most recent version of this article at:

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

Subscriptions

Reprints and

[email protected] 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/49/14/3891To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



an epitope common to gangliosides 0-acetyl - cancer research

Documents