CA1210689A - Method for detecting the presence of g.sub.d.sub.3 ganglioside - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE

Determining the presence of GD3 disialoganglioside in a test sample by reacting antibody R24 (AbR24) with the test sample and determining the extent of the reaction.

Description

6~
~ BACKGROUND

Ih ~7 ~15 e, We have previously described a-m~s~ IgG3 monoclonal anti~ody (AbR24) with a high degree of serological specificity for cultured human melanoma cells (1). All melanoma cell lines examined and two astrocytomas were positive for the heatstable cell surface antigen detected by this antibody. Although choroidal melanocytes and brain had low levels of the antigen, a wide variety of other cells and tissues were unreactive.
Three other monoclonal antibodies (Abs C5, I24, and Kg), having a similar restricted specificity, we~re derived from the same fusion. These antibodies showed the same strong reactivity with~melanomas and lack of reactivity with epithelial cells, but had a slightly wider specificity range in that they also reacted weakly with MOLT-4 (a T-cell line), leukocytes and some fetal tissues.

The antigen detected by Ab R24 is identified herein as G~3_ a previously characterized disialoganglioside. In compa-rison with other cells and tissues, melanomas have high levels of GD3. Thus, these antibodi~s are useful in determining whether a tissue sample is a melanoma or not. This is parti-cularly important for characterizing lesions. ~hese antibodies can also be used in determining concentrations of GD3 in serum, plasma, urine or other body fluids. This may aid in the early dia~nosis of melanoma and possibly of other disorders whcre there are elevated glycolipid levels.

-~2~ 6~

SUk~S~RY
Mouse monoclonal antibody AbR24 (Dippold et Q~. Proc. Natl Acad. Sci.
77:6114-6113 1980~ has a high degree of specificity for human melanoma cells wllen tested on viable cultured cells using the PA-~n~.~ serological assay. The antigen detected by this antibody has been isolated from melanoma cells and shown to be GD3 ganglioside by compositional and partial structural analysis and by comparison with authentic GD3 by tl~in layer cl-romatography (T~C). AbR24 reacts with authentic G~3 but not with any otller ~allglioside tested. Using TLC and reactivity with AbR24 a wide range oE cells and tissues ~JS e~:amined for the presence of GD3. A new serologic~l ass;ly tel~ ed glycolipid-mediated imroune adherence (G~IIA) was devised for assayin~ tl~e reacti~ity of AbR24 with gangliosides. kSelanomas (cultured cells or tumor tissue) were shown to have GD3 and G~3 as ma~or ~angliosides. Other cells and tissurs e~aroined also con-tained CD3 but usually only in lo~ amounts. ~Iclalomas (and t~l.T-4 a T-cell line) were characterized by a simplified ganglioside profi~e witll GD3 and G~l3 as major components. The apparent dlscrepancy betwcen the uhi(l!liL~u~ presence of GD3 and the serological speciEicity of ~bl~74 fo melanoma cells c n be e~plained in terms of localizatioll and concentratioll of ~3 il- diff~lenL ceIIs.

.~ . .

?~

BRI~F DESCRIPTION OF DFUAWINGS

IGURE 1. Tlme course for the reexpression o AbR24-reactive antigen on SK-~EL-28 cells after neuraminidase treatment. Assay: PA-~-~.
IGURE 2. Locallzatlon of AbR2b-reactive glycolipid on thin layer chromato-graphy using glycolipid-mediated immune adherence (G~IlA) assay.
Acidic glycoliplds from SK-~IEL-28 cells were separated by TI.C in solvent 1. Silica gel bands (1 cm wide) were scraped from the plate, extracted with C:~t (1.2) and assayed Eor antiEens by G~IIA
as described in the te~t., IGUR~ 3. Thin layer chromatography of acidlc glycolipid fractions from a number of cell lines and tissues. Lane 1: GI13; 2: GD3; 3: G~
4: 'SK-IIEL-28 melanoma cell line; 5: AhR~4-reactive antigen iso-lated from SK-~IEL-28; 6: SK-~L-37 melanoma cell line; 7: SK-rI~L-64 melanoma cell llne; 8: ~leI~o; 9- SK-~IEL-13 melanoma cell line;
10: melanoma (surgical specimenj; 11: IIOI.T~ cell line; 12: mouse eye; 13: SK-RC-7 renal cancer cell line; l4: ~du]t human brain.
Gangliosides were separated in solvent 1 and visualize~ with resorcinol-HCl.
IGUR 4. Densiometric tracings of thin layer chromatograms of gangliosides from melanomas and other cells. A: S~-IIEL-2S melanoma cell line;
B: SK-~IEL-37 melanoma cell line; C: SK-iiEL-13 melanoma cell line;
D: melanoma (surglcal specimen); E: ad~ lullllall brain; ~: Raji B-cell line; G: ~IOLT-4 cell llne; Il: ~-I`C-7 rennl cell line.
The amount of GD3, as % of total ganglioside fraction, was calculated from the areas ot the pe.~k.s aIld ls indicaIelI In e ncll panel.
JGURE 5. lnhibition o~ reactivity of AbR"4 witII SR-~IEL-)S melnIlonla cells by acldic glycolipid fractlons from a variety of cell llnes and tissuus.
Assay: rA-~IHA. ~ : AbR24 control;'~' : A~ t hIlm;ln sI)leen;

. .
~,\,r _4_ ~2~qt~

~: adult human liver; O: teleost eye; ~ : SK-RC-7 renal cancer cell line; ~: adult hnman brain; O : ~leWo melanoma cell line;
: S~-MEL-29 melanoma cell line; V: S~-~L-37 melanoma cell line;
~: mouse eye; n: ~IOLT-4 T-cell line; ~: ~elanoma ~surgical speci~en); a: SK-MEL-28 melanoma cell line.
IGURE 6. Glycollpld-~ed$ated imm~ne adllerence (C~IA) assay using AbR24.
Well A: AbR24-reactlve glyeolipid lsolated from SK-MEL-28 melanoma cell line; well B: GD3 gan~lioside; well C: no 6anglioside;
well D: G~l2 and G~3 gangliosi(le mi~ture. Antibody: Ab~26 (l:lOOO).
IGU~E 7. Detection of GD3 ganglioside by AbR~4 ln GMIA assays. AbR24 dllutions are lndicated in the figure.
IGURE 8. Detection of GD3 gangliosld2 on TLC plates by reactivity with AbR24 and l25I-Protein A. R$ght side: gangliosides visuallzed with resorcinol-llCl reagent; leEt side: gangliosides reactin~ with AbR24 and ~51-Protein A. Lane l: AbR26-reactive ganglioside;
Lane '2: sangliosides extracted Erom adult human brain. Solvent 2.
IGURE 9. PrQposed pathways for the biosynthesis o~ gangliosides Imodified after Y~ and Ando (32)], .

DETAILED DESCRIPTION

Tissue Culture _. .

For derivations and culture of melanoma and other cells see references 1-4. Normal and malignant tissue was obtained from surgical or postmortem specimens.

Monoclonal antibodies Mouse monoclonal antibodies AbR2~, AbC5, ~bI12, and AbNg have been previously described (1). AbR2~ and A~C5 are IgG3 antibodies and AbI12 and AbNg are IgG2b and IgGl anti-bodies, respectively.

Glycolipids GD3 was obtained from Dr. Y-T. Li, Tulane University, New Orleans i5). GM3 and GM2 were obtained from Drs. 5. Kundu and ~.M. Marcus, Baylor University, Hous~on, TX. GMl, ~Dla~ G~
were purchased from Supelco, Inc., Bellefon~e, PA. Lactosyl-eer~nide was purchased froTn Glycolipid Biochemical Cv., Birmingham, 6 . _ , .. ..

Serologlcal assays fcor melanoma cell surface anti~ens Reactlvity of AbR24 and AbC5 with cell surface antigens of melanoma cells was determined wlth cultured cells growlng in the wells of ~71crotest plates (Falcon 3034~ using a red cell rosetting method (3) in which indicator cells are In~man O rPd cells (RBC) to which StaphyZococclls aureus protein A is coDju~ated (PA-~). AbI12 and AbN9 were assayed using a modification oE this method in which rabbit antl-mouse Ig-con~ugated indicator cells were used (IgG-~A).
~nzyme treatment Melano~la cells growing as monolayers in microtest plates as described above were washed witll Hank's balanced salt solutlon lHBss~ llicrobioloeical Associates) and then treated with Yi~ro c7w~erae neuraminidase (Calbiochem) or e-~alacLosidase isi~ma. Type VIIj using 1 U/well in lO ~l of II~S.S. After lncubatlon for 1 hr at 37~, the cells were washed four times ~ith PBS - ~X
ga~Qa globulin ~GG)-free F~S and assayed for reactivity w{th antibody using the PA- or lgG-~nlA assays.

Isolation of glycoliplds .
Glycolipids were lsolated initially by a modification of the method of Saito and llakomori (6), and separated into neutral and acidic fractions by DE~E-Sephadex chromatograplly (7). Acidic ~lycolipids (gan~liosides) were sub-sequently isolated directly from chloroform-metllanol (C-~l) extracts by DE~E-Sephadex chro~ato~raplly and alkaline llydrolysis (7). I.rie~ly~ cclls wcre homo-~enized in C-~ nd after filtrntion were rc-exLraclcd wltil C-~l (l:lj.

After evaporatin~ and redissolving the extract in C-~ ), it is filtered, e~/aporated and dlalyzed agalnst distilled ice water for 24 hours in the cold.
After dialysls, samplQs were evaporated, dissolved in C~ H2O (30:60:B) and applied to a DEAL-Sephadex column (equ~librated wl~h C~ 0.8~1 Na acetate);
(30:60:8)~ After washing the column with C~ O ~30:60:~),acld1c liplds were eluted w~th C-M-0.8~ Na acetate ~30:60:8), evaporated and dialyzed as beore.
The acidic fraction was then ~lydrolyzed wlth 0.1 N Na0H in methanol for 3 hours at 37C, dialyzed against cold water (48 hours), evaporated, and dissolved in C~ 4:1). The solution was applied to a Biosil A column which had previously been washed with C-~ ~4:1). Af~er eluting impurities with C-M ~4:1), gangliosides were eluted with C~ 2).
Thin layer chromacography tTLC) Silica gel plates ~ediplates, Fisher Scientific Co.) were activated by heating a~ i200C for 1 hour. Solvents used for developin~ chromatograms were n-propanol-NH4OH-H~0, 60:9.5~11.5 (solvent 1) as ln ref. 8 and chloroform:
methanol:2.5N NH4OH, 60:40:9 (solvent ~. Once the solvent had migrated 1~ cm from the origin, the plate was removed and air-dried for 12-15 minutes at 110-120C, cooled to room temperature and sprayed wlth resorcinol-~Cl (9).
For preparative analysls, plates were dried at room temperature in an air flow hood for 2-3 hours~ Bands were visualized with iodine vapor, ou~lined and silica gel scraped rom the plate. The gel was then extracted twice with 40 ml of G-~ 2O (50:50 15~, with à small amount of Dowex 50 (Na ~. Ihe suspension was centrifu6ed at 1~00 rpm for 15 minutcs an~ tlle solution filtered, evayorated, redlssolved ln C-~l (4:1) and applled to a ~iosll A colunul as describ~d a~ove.
lmpurities were eluted witll C-~ (4:1) .~nd adsorbcd gan61iosides wer~ then eluted with C-M (1~
Carbohydrate analysis Lipld-bound sialic acid in cell pellets was dctermined on C:M (2:1 and 1:2) .. . ..

extracts after hydrolysis in O.lN HCl at 80 for 1 hour as described by l~arren (10). Sugars were analyzed after methanolysis (methanolic 1.01~ IICl at 100 for 16 hour5) as thelr 0-trifluoroacetates {11~; N-acetylneuraminic acid was identified by the sanne procedure aEter methanolysis in l.ON liCl at 80 for 1.0 hour.
Serolo~ical assays for ~lycolipids (iJ Passive hen~laglutlnation assay (cf. 1~) Glycolipids (6 ~g sialic acid) were dissolved, aliquoted into tubes (10 x 75 mmj and dried in a dessicator witl) P~05 in vacuo. ~o each tube, '00 ~1 of PBS was added, the sides of the tube scraped and the solutions soni-cated for 15 min at 50C. After transfer to a lar~er tube, 0.8 ml of PBS was added. The glycolipid solutlon was added slowly in a dropwise filshion to a 2%
suspension of human 0-RBC in PB5. A~ter 1 hour at 37C, with one r~ixing after 30 minutes, tl~e cells were washed twice with PBS ~lZ ml -each wash~. 2eactivity was tested by ~ixing a suspension of the treated RBC and appropriately diluted AbR~4 (50 ~1 eacll) in microtlter plates. After 1-' I-ours at 4C, the a~luti-nation reactions were scored visually.
(iij Antibody inhibition assay Glycollplds (6 ~g slalic acid),dissolved in C~ j,were ali-quoted into tubes (6 x 50 mm) and drled as in the previous assay. AbR24 (l:' x 104) was added (30 ~1~ and the tubes were vortexed, incubaCcd for 30 mlnutes at room temperature, and then for 30 minutes at 4C. Tubes were centrlfu~ed for Z0 minutes at 2000 rpm and thc sllpernatants removed and serially d11ute~. These samples were imme~liltcly LransEerred to EormDldcll~de-fixcd S~IEL-2B tar~ct cclls. IThc rormalll~l)yd(: Fl~iltion was c;lrric/l ouC by treacin~ cells ~rowin~ in the wells of microtesL plates ;Falcon 3034) witll 0. 33% IICIIO in P~. Tl~is treatment does not alter re~ctivlty witl~ /~bR,4 and provides a store of readily available source of tarnet cellsl. ~ntibody .. . _ . . .

reactivlty was detected wlth the P~ ~ assay. ~nabsorbed antlbody served as a posltive control.
~lil) Glycolipld-mediated lmmune adherence assay (GtllA~
A 601ution of glycolipids in 95% ethanol was added to the wells of mlcrotest plates (Falcon 3034; 10 ~1 per well) and the plates were dried in a dessicator in vacuo with P205 f or 45 mlnutes. Approxi~ately 100 n~
of lipld-bound slalic acid was found to be the optimal amount for efficient adsorption and ~aximal reactivity with antibody. Wells were then washed three ti~es with PBS- 2~ GG-free FBS (10 r~/wash), and the plates blotted with gauze.
Diluted antibody (in PDS with 5~ GG-free F~S) was ~dded to the wells and incubated for 45 minutes at room temperature. ~lates were blotted, washed four times with PBS- 2% GG-free FBS, and blotted a~ain.
~en ~1 of a 0,2~ suspension of Prote~n A-con~ugated o-~C ~tere added ~o the wells. The plates were incubated for 30 minutes at room temperature.
After blottlng, the plates were washed twice with PB~ 2% GG-free FBS, blotted once a~ain and read under the light microscope. Reactions were scored according ~o the proportion of the well which was covered by red cells. A test was read as negative when wel~s showed no adhering cells or only a thin rin~ of cells around tlle perimeter.
(iv) Detection of serologically-reactive ~lycolipid after separation by thln layer chromatography.
Serolo~ical reactivity of glycoli~ids separ~ted by thin layer chromatography was tested usin~ a modiication oE the metllod of ~la~nani ~t a~.
(13).in which 1 51-Protein ~ was used to detect tllc bound antibody. ALtcr-chrom.~to~raphy in sclvents 1 or 2 thin l~yer sllee~s ~lcre washcd 1l) I'ùS buffer containin~ 1% polyvinylpyrolidone and ~reated witll AbR74 (1:150n) ~or 6 hours ~t ~C.
AEter washin~ in P~S, the plate was treated with 125I-Proteln A (10 llCl/~
7 x 10~ cpm/r~), prepared accordin~ to tlle procc~ure of Hunter and Greenwood (14).

. .

--10~

~2~

After standing for 12 hours at 4'C, the plate was washed in PB~ air-dried anrl exposed to X-Omat R film wlth a Cronex intenslfier screen for 2-6 hours.
RESULTS
teration of Ab~ serological reactivi~y and klnetics of anti~en restitution after neul~nin idase treatment o~ SK-~EL-28 After treatment with neuraminidase (Vibr~o cho~erae), SK-~IEL-28 melanoma cells no l~nger reacted with ~bR24 ln P~-MI~ assays (Table I). Reactivity witl ~bCs 1an antibody with a serological specificity simllar to that of AbR24 (1)~
was also lost. Reactlvity with AbNg and ~bI12 whlch recognize serologically unrelated determinants on glycoproteins of SK-~IEL-28 was unaffected by neura-minidase, Enzy~e-treated cells did not show non--specific reactivlty with either Protein ~-or with anti-mouse Ig-lndicator cells. B-Galactosidase had no detect-able effect on the reactivity of SK-~EL-28 cells with ~bR24 or ~bCs (Table 1).
These results show that sialic acid constitutes an important part of the antl-~enic determinant recognized by antibodies ~bR24 and AbC5.
Serological reactivi~y of ~bR24 with SK-~IEL-)8 remained undetectable for 30 minutes after neuraminidase was removed and replaced with ~E~I-FBS. Contlnued incubation in this medium at 37C resulted in a partial return of AbR24 reactiv-ity after 2 hours and complete recovery o serolo~ical reactivity after 22 hours(Fi~ure l)o Isolation of Ab~24-reactive antigen from Sl~-tEL-28 melanoma cells and its identificatlon as Gn~ ganglioside Glycolipids were isolated from cultured melalloma cells (5~-)1EJL-28) by cl~loroform-meLIIanol (C-~l) extraction and ~loris3l cllromato~raptly o[ their acetates as dcscri~cd by Salto and llakomori (5) ~nd th~ ~lyc~ )id pL'~ la-tion was fractionated into neutral and acldic compollents hy Dr:,~E-SephadeY.
chromato6raplly. Inhibitory actlvity ~ainst AbR~4 nllti~ody ~nssayed . . .

with PA-~) was found to re6ide entlrely in the acidlc glycolipid fractions.
In subsequent exper~ments, acldic glycolipids from S~-~L-28 cells were isolated directly by fractionatlng the C-M extract on DEA~-Sephadex (6) and eliminatlng contaminating pllosphollpids by alksline hydrolysis. Indivldual gangliosides ln this mixture were isolated by preparative thin layer chromato-~raphy in solvent 1 (8). By scraplng a series of silica gel bands from the plates and extracting the glycolipids, the antigenic activity was located in the ~a~or acidic gly~olipid band whi~h migrated just above G~ and GDla ~Figure 2). In the data presented in Figure 2, the antigenlc activity of fractions was measured by the GMIA assay. Similar results were obtained by antibody inhibi~ion tests usin~ the P~-r~ assay with AbR~4 and S~-]IEL-~8 target cells~.
The isolated AbR24-reactive glycolipid was identified as GD3 [NAN~(2-~8)N~NA
(2~3)Gal~ 4)Glc-ceramide] by ~he following criteria: (i) carbol7ydrate analysis of the purlied &lycolipid showed that ;t contained glucose, ~alactose and N-acetylneuraminic acid in a ratio of 1.ool.o~:2.~ ithonly a trace (< 0.1) of hexosamine,- ~ii) partial hydrolysis of the ganglioside with Vibrio cholerae ~euraminidase (3 hours at 37C) resulted in the formation of two co~nponents comigrating on thin layer chromatograms wlth G~l3 and lac~osylceramlde, Slii) the purified melannma glycollpid comigrated with authentic GD3 in thin layer chroma-tozraphy (Figure 3) and (iv) ~bR24 reacted with authentic GD3, but not with any of the other standard gangliosides tested (see below~.

Distribution of GD3 in melanoma and non-melanon-a ce]l lines, anu in normal and mali~nant tissues (i~ Thin layer chromi Lo~.ra~ c pa~l:orllJ: Or 1,;1nl]il-!iil3C'i r~O~ v.,rl(llls sources Tutal ~anglioside fractlons were prepar~d Lrom a lar~e varlety ~rl -12-J;~

of tumor cell lines, fresh tumors and normal tissues. ~en these e~tracts were fractionated by TLC and the gangliosides detected using the resorcinol reagent, it became evident that melanomas have a characteristic pattern of gangliosides.
In all the melanoma cell lines examined, glycoliplds comigrating w$th GD3 and G~13 were prominent acidic glycolipids, with ~D3 being the ma~or component in many of these cell lines (Figure 3 and Figure 4). GD3 was also a prominent ganglioside in extracts of mouse eye and bovine choroid. With the exception of ~IOLT-4 (a T-cell line)J none of ~he other cells or tissues had GD3 as the ~ajor component. Extracts of fresh melanoma tumors gave gang]ioside patterns resembl-ing SK-~FL 28, with GD3 and G~3 predominating (Figure 3). Most melanoma cell lines gave thls 5imp~ lfied pattern, but some showed a more complex profile in whlch higher gangllosldes were detected in appreciable amounts (Flgures 3 and 4). Gv3 constituted 18-63~ of the total gan~lloside fraction in the melanoma cell llnes examined (Figure 4). ~lost mclalloma cell lines and s~ecimens had values in the 30-5~`~ ran~e. Tl~ese vnlues compare~ with 7Z in adult hurlan brain, 9~ in a renal cancer cell line (SK-RO 7), 11% in RAJI cells (a Burkitt~s~lymphoma) and 33% in ~IOLT-4 cells (Fi~ure 4). In terms of the serological reactivity of AbR24, it is ~mportant to note that melanomas, in addition to having higl)er proportions of GD3 in tl~eir glycolipid fraction, also have higher total ganglioside levels. This is evident from a determinatlon of the levels of lipid-bound sialic acid in a number of cell lines. In melanomas the values ranged from 0.039-0.063 ~ molelO.l ml cells (determined on 9 lines).
RAJI, ~IOLT-4 and ~enal cancer cells (3 lines) had lipid-bound sialic acid values of 0.011 ~ 00003t 0.013 ~ 0.006 and 0.025 - 0.029 ~I mole/O.l ml cells, respectively.
(ii) Dctection of GD3 in cell llnes nlld tl~snr~ usin~ hh~4 nlltlbody GD3 levels ln a lar~e variety of cel1s and tissues here estimated usin~ R24 antibody. Four assay methods ~ere used: (a) passive hemagglutination, ~2~ i&~

tb) antibody inhibitlo~ c) a new method, G~IA, devised to combine the simpllcity of the ~LHA method with the ability of glycolipids to adsorb to plastic.and (d) a method using 125I-Protein A to detect AbR24 reacting with GD3 on TLC chrom~tograms. The sensitivity of the assays varies consider~bly;
the passive hemagglutinatlon assay is the least sensitive and the 125I-PA
method the most sensitive (Table II).
~ sing the least sensitive detection method (passive hema~glu-tination), GD3 could be detected in extracts of melanoma cPll lines and ~elanoma tissue, bur. not in other sources ~Table II). ~lore sensitive assays (inhibition of PA-~n~ and G~ methodsj sho\~ed that GD3 was detectable in a wider range of cells (bovine choroid, mouse eye, fetal and adult hu~an lung, RAJI B-cell llne, ~IOLT-4 T-cell line, RT-4 bladder cancer cell line and AJ astrocytoma cell line). A typical inhibition experiment is presented ~n Figure 5 and ~he data are summarized in Table II. Usin~ the G~lIA method lt was found tha~ wells coated with R~4-reactive glycolipids from melanoma (Figure 6A) or authentic GD3 save stron~ly positive reartions (Figure 6B); some quantitatlve data on this reaction are sho~n in Figure 7. Other purified glycolipids (G~ GDl~. G~13 and C~l2) were unreactive in this assay (Table II and Figure 6D~. AbR24 added alone was also unreactive ~ ure 6C).
Application o~ this method to acidic glycolipids e~tracted from other cells ~ave approxlmately the same results as inhibltion assays ~Iable I13. In contrast to the restricte~ distribution ~ Gl)~ deLc~Lmilled by these methods, the 125I-Proteln ~ method detected GD3 in all the cclls Dnd tissues examined (T~ble Il). That tl~e ~bl~4-re~ctivc con-l-onellt slctocLed in the~e tissues and cells ~35 in fact (:D3 was indicn~e(l by its uo~ r;l~ion :iLII ;luLIIcnLlc (;D3 ~ln t-~o solvent systems~, and by the findin~ tln~t a~lotllcr proL~in /~-bindin~
monoclonal antibody (~ l, detectinG an unrelated ~Iycorroteill spccificity, ~ -14-;

was unreactive.
DISCUSSION
Mou5e monoclonal antibody R24, which shows a high degree oE serological specificity for cell surface antigens of melanoma cells, recognizes a disialoganglioside - GD3. Past studies have shown that antibodies to gangliosides have been difficult to raise (15). This may have to do with the fact that most gangliosides are constituents o~ the species being immunized and also, because in situ sialidase-activity may destroy ganglioside immunogenicity (16). In this regard, it might be signi~icant that the mouse from which AbR24 was developed had been extensively immunized over a period o~ six months with melanoma cells (SK-MEL-28) having a very high GD3 content.
SK-ME~-28 has been deposited at the ~merican Type Culture Collection (ATCC) 1230 Parklawn Drive, Rockville, Maryland 20B52 as HTB 72. Two other monoclonal antibodies recognizing glangliosides have recentl~ been described (17,18). One reacts specifically with chicken neuronal cells and is directed against one of the higher glangliosides present in the G~ fra~tion (17); the second is directed against human colon carcinoma and recognizes an as yet uncharacte-rized monosialoganglioside (18).

We have shown that GD3 is a prominent ganglioside in cultured melanoma cells and in melanoma tissue. When compared with other cells, melanoma cells also possess relatively hi~h total ganglioside levels. As shown by others, GD3 is present in small amounts in most mammalian tissues, but it is a major ganglioside in the retina, where it comprises between 30-40~
of the gangliosides (19). In adult human brain, GD3 represents about ~-10% o~ ~he total ganglioside content (19). Levels of ~2~Y~

GD3 may be higher in fetal brain, considering that in fetal rat brain (15-17 days gestation) GD3 represents about 50% of the total ganglioside content, falling rapidly to about 10% by day 20 (20). Portoukalian and coworkers (21) have also reported that GD3 identified by TLC and carbohydrate analysis, i5 a major constituent of melanomas. They showed that the proportion of GD3 varied from 31~0~ to 57.;2% of the ganglioside~

- 15a -fraction in the four di~ferent melanoma specimens examined. From ~hese resulta, as well as our owrn analysls, one can conclude that CD3 g~nglioside is a yroml-nent component of malir~nant melanoma. Whether normal melanocytes have high levels of GD3 is at present unclear. Normal choroidal melanocytes show weak reactiYity with AbR24 ln direct serolo~ical tests (titer 1:100) as comp~lred to the strong reactivity of melanoma cells (titer of 1:5 x 10~ - 105) (1). With the recent development of a method for culturing skin melanocytes (22)J.it will now be possible to m~ke a direct comparison of tlle ~D3 contcnt of melanocytes and melanomas. Although a precise biological function Eor GD3 remains to be determined, lt has been suggested that GD3 has a role in serotonin binding ~23,24).
In examining the TLC patterns of the ~angliosides isolated from different melanoma cell lines, we noticed considerable vnriation in the proportion of thc various ~angliosides. In most cell lines GD3 and G~l3 were the predominant ~angllosides (Figures 3 and 4). A few melanoma cell lines showed ~ more com-ple~ pattern with C~2 and some hi~her gangliosidcs bcin~ better reprcseneed;
whether thesc differences in ganglioside profiles corrtla~e ~ h biological characteristics (c.~. differentiation state) of the tumor needs to be determined.
In genera~ melano~as e~hibit a distlnctive gan~llosldc ~rofile. or the otller cells and tlssucs examined, only tlle T-cell line IIOI.T-4 sho~cd a similar pro-file, nnd tllls may be ~nother example of anti6ens shared by T-cells nnd cells of neuroectodermal ori~in ~ Thy~ 5). ~all611Osides derivetl from bovinc choroid and mouse eye llad more complex p;lLtcr-l6, ~ 3 hcinr. ~nl)~ one of three or four promincnt components.
It i6 vcry cvident from tllc nn;ll;ucis -r cxt rn~ ])colipitl.q tll;lL thc pre6encc of ~D3 ~anr,lioslde is by no mcans rc.cLrLcl.d lo mclalltma ccll~ - iL
is ublquitous. Yet usln~ direct serolor,ical assays for ccll surface alltigcllc, only melanor.)as, choroid;ll melanocytcq, and n.stroc)tomas ucle icactive ~ith 1~ ' with AbR~4 (l). Even using sensitive ah~orption tests, only normal brnin of other cells and tissues tested absorbed AbR24. A nuTnber of explar1ations for ~he apparent discrepancy between ~he serological finding and the biochemical data presented here can be suggested. First, lt is possible that GD3 is not a cell surface constituent of most non-melanoma cells. It is well established that GD3 is a biosynthetic precursor of other ganGliosides (Figure 9) and would therefore be located mainly within the cell, probably in the Golgi apparatus where the glycosyl transferases responsible for glycolipid synthesis are found ('6,27). As our blochemical studies were carrled out on whole cell or tissues the results are certainly compatible with this exp]anation. Another possibility is that GD3 is present at the cell surface of R24- negative cells but is not avallable for~reaction wlth ~ntibody. This phenonenon has been found with other cell tnembrane glycolipids ~ loboside is a Major glycolipid o~ erythro-cyte membrane but erythrocytes react only weakly ~ith anti-globoside antibody ~28). It is also possible, of ccurse, that GD3 is not cxpressed on the surface of most non-melanoma cells in amounts that are detcctable by the serological tests used. Tt is important to note tllat the cell types which reacted with AbR24 in both direct and abscrption tests have both a hl~h lipiù-bound sialic acid content and have GD3 as a prominent gtt1glioside.
1~hat ~ight be the mechanism of the accumulation of GD~ and G113 in melanoma cells? ~ne ~ossible explana~ion is that n~clanon1a cclls have low ]evcl.q of ~-acetylgalactosamlnyl transferase~s~ whlc11 ~ould result in the accun1ulation oE the normal substrates for the enzymets) i.e. G~3 al~d Gn3 (Fi~ure 9). ln bovine thyroid, a sin~le N-acetyl~al~ctosamine-trans~c1ase is thought to act on botl- ~D3 and ~13 to ~orm GD~ and ~?1'' ('7) ~n~ VL~IS or t1~is ~n~vm~
in ~elanomas coulLI explain the ~an~liosl~1e pattcrn wc observLd. Among ot1~er possible e~planations are that n1cl;tnnma~ h;lve hi611 ].eVOIY oE r;-;l-ace~yl~al;-l Lo-saminidase which would result in increased degradation of G;~, and G~2 or that rnelanomas have elevated levels of certaln sialytr~nsferases, resulting in increased synthesis of CD3 and GM3. It i5 si~nificallt in this re~ard that melanoma patlerlts have i.ncreased serum sialyltransferase levels (Z9)~ ~nzyme levels in tumor t.issue have not yet been studied, althou~h the fact that the glycoproteins of human melanoma cell lines have increased sialylatlon as com-pared to the ~lycoproteins of other cell types (30) sug~ests increased activity of thls enzy~e ln melanoma, Abbreviation Used in_~his Application TLC: thin layer cl~romatography ~A: mixed hPma~lutination assay C-~l:
chloroform-me~hanol; FBS: fetal bovlne serum; NANA: N-acetylneuraminic acid;
Gal: D-galactose; Glc: D-glucose; GalNAc: N-acetyl-D-galactosamine; Cer:
ceramide; G~ Gal 1-3 GalNAc 1+4 ~Gal~3~2 NANA] 1'4 Glc-Cer; G~13: NANA

2~3 ~&al 1~4 Glc-Cer; G~3: NANA 2+8 NANA 2i3 RGal 1~4 Glc-Cer; GDl : NANA
2+3 ~Gal 1+3 GalNAc ~ 1+4 Gal ~3l2 N~NA] Glc-Cer; G~2: ~GaLNAc 1'4 ~Gal ~3~2 NANA) Glc-Cer, GTla: NANA 2-~8 NANA 2-~3 ~Gal 1-~3 GalNAc ~1~4 Gal ~3+2 N~NA~ Glc-Cer.
[Nomcnclature of Svennerholm (31)].

REFERENCES
1. Dippold, W,G., Lloyd, K,O., Li, L.T.C., Ikeda, H., Oettgen, }I.F.
and Old9 L.J. ~1980) Cell surface antigens o~ human malignant melanoma: Definition of six anti~,enlc systems with monoclonal antibodies. Pr_c. Ilatl. Acad. Sci. US~ 77:6114.
2. Carey, T.E., Takahashi, T., Resnick, L.~., Oettgen7 Il.F. and Old, L.J.
(1976) Cell surface antigens of human rnalignant melanoma: mixed hemadsorption assays for humoral ~rnmunity to cultured autologous ~nelanorna cells. Pro~. Natl. Acad. Sci. USA 73:3278.
3r Pfreundschuh, M., Shiku, H., Takahashi, T., Ueda, R., ~ansohoff, J., Oettgen, H.F, and Old, L.J. (1978) Serological analysis of cell surface ant~gens of mallgnant brain tumors. P;~ Natl~ .~c~d~ Scl~ IJSA
75:51'2,~
4. Ueda, ~., Shiku, H., Pfreundschuh, ~1~, Takahashi, 1'., Li, L.T.C., Whitmore, W.F., Oet~gen, H.F. and Old, L.J. (1979) Cell surface antigens of human renal cancers defined by autologous typing.
J. Expl. ~5edO 150:564.
5. Itoh, T., Li, Y~T., Li, S-C. and Yu~ RoK~ (19~1j Isolation and characteri-zation of a novel monosialosylpentahexo~yl ceramide from Tay-Sachs brain.
J. Biol. Chem. 2 :105.
6. Saito, T. and Ha~omori, S. (1971) Quantltati~e isolation o~ total -glycolipids rom animal cells. J. Lipld Res. 12:257.
7. Yu, R.K. and Ledeen, ~.W. (1972) Gan~liosides of lluman bovine and rabbit ~ sma. JO Li~id Rcs. 13:~80.
8. ~Jatanabe, K., Powell, ~I.E. and llalcorn~rl, S. (1979) lsolation and cllaracterization of gangliosides witl~ ~ new 5ialosyl linlcagc and ~2~q~

core structure. II. Gangliosides of human erythrocyte membranes.

J Biol, Chem. 254:~223.
_ _ 9. Svenllerholm, L. ~1957) Quant'itative estimation of sialic acicls. II. A
colorimetric resorcinol-hydrochloric acid method. _iochim. Biophys.
Acta 24:604 _ _ .
10. Warren, L. (1963) Thlobarbituric acid assay of sialic acids. ~lethods in Enzymol. 6:463.
11. Zanetta, J.P., Breckenridge, W.C. and 'Vincendon, G. (1972~ Analysis of monosaccharides of the 0-methylglycosides as trifluoroacetate derivatives.
J. Chromato~r. 69:291.
12. Yokoyama, ~1., Trams, E.G. and Brady, RØ (1963) In~nunochemical studies with gangliosides. J. Immunol. 90:372.
13. M~gnani, J~L,, Smlth, D.F~ and Ginsburg, V. (1980) Detection of ganglio-sides that bind cholera eoxin: direct bindin~ of 1 5I-labeled toxin to thin layer chromato~rams. _nal. ~incllem. 109:399.
~4. Hunter, ~ S. and Greenwood, F.C. (196')) Preparation of iocline-131 labeled 'human ~rowth hormone of high specific activity. Nature 194:495.
15. Rapport, lt.~l. and Graf, L. (1969) Immunochemical reactions of lipids.
Pro&. ALlergy 13:273-331 (1969).
16. Kundu, S.K., ~larcus, D.~l. and Veh, R.l~. ~l980) Preparation and properties of antibodies to GD3 and G~ an~liosides. J. I~lellrocllcm. 34:184.
17. E;isenbarth, G.S., Walsh, F.S. and ~'irenber~, Il. (]979~ ilonoclonal antibody to a plasma membrane anti~en of neurons. I'roc. Natl. Acad. ~ci. US/~
76 4913.
18. ~laE~nani, J.L., Brockhaus, ~l., Slllilh9 ~ sl)ur~ ~ V.~ Ula~ y~uk, 11.~
~litchell, K.F., Steplewski, Z. and 7;opL-!wskl, Il. (19Sl) J~ mnnclslalo~n~ lioside is a monoclonal antibody-defined anti~en c~ c~lon carcinoma. Science '1'':55.

_ _ l9o Urban, P.F., Harth, S., Freysz, L. and Dr~yfus, H. Brain and retinal gan~lioside composition from different specie~ by TLC and IIPTLC. in Structure and Function of Gangllosides. Adv. Exp. ~_d_ Biol., ed.
L~ Svennerholm (Plenum, New York), Vol. 125,~ p. 149 157, 1980.
200 Irwin, L.N., tlichael, D.B. and Irwin, C.C. tl980) Gan~lioside patterns of fetal rat and mouse brain. _ Neurocllem. 34:1527.
2l. Portoukalian, J., Zwingelstein, G. and Dore, J. (1979) Lipid composition of human malignant tnelanoma tc~mors at various levels of malignant growth.
Eur. J. Biochem. 94:19.
-22. ~islnger, ~1. and ~larko, 0. Selective proli~eration of normal humanmelanocytes in vitro in the presence o~ phorbol cster and cholera toxin. Proc. Natl._Acad. Sci. US~ (in press).
23. Wooley, D;W. and Gomml, B.W. ~1965) Serotonln receptors VlI. ~ctivities of various pure gangliosides as rec~ptors. Proc. Natl. Acad. Sci. US~
53:959.
24. Tan~lr, H., Brunner, W., Casper, D. and Rapport, ~I.il. (19S0) Enl~anccment by gangllDsldes of binding o~ serotonin to selotonin bindin~ proteins.
J Neurochem. 34:1719.
_ __ 25. Reif, A.E. and Allen, J.~l.V. tl964) Tllè ~R tllymic antizen and lts distribution in leukemias and nervous tissue. ~l E~p. ~led. 120:413.
76. Keenan, T.W., i~lorre, D.J. and Basu, S. (1975) Gan~lioside bio~ynthesis~
Concent~atlon of ~lycol~l;in~olipid ~lycos;l trnns~crasc~ in Goi~i apparatus from rat liver. J. Biol. Cllem. 249:310.

_ ~7. Pucuszl;a, 1., Duf~ard, RØ, Nis1~imll~a, r~.N., ~rndy, ~.r. al~d Fisllman, r.ll.

(1979~ Dlosyntl~csis ot l)ovlllc tl~yroi(l ~;~n~llo~ s. .I._';ul. ~ m.

253:5839.

28. Il~komor~ 9 S. (1973) Glycollpids of tumoJ ccll ~cm~rane. ~dv. Callcer Rcs. 18:'65.

..... .
-21~

~2~
..

79~ Silver, H,K.B., Karimj K.A., Archlbald, E.L. an~ Salinas, F.A. (1979) Serum sialic acid and sialytransferase as monitors of tumor burden in malignant melanoma patients. Cancer Res. 39:5036.
30. Lloyd, KØ, Travassos, L.R., Talcahashi, T. and Old, L~J. (1979) Cell surface ~lycoproteins of human tumor cell lines: unusual characteristics of malignant melanoma. J. Natl. Cancer Inst. 63:623 31. Svennerholm, L. (1963) Chromatographic separation of human brain gangliosidesO J. Neurochern. 10:613.
32. Yu, R.K. and Ando, S. (1980J Structures of some new ~angliosides of fish brain in Structure and Function of Ganglio~ide (ed. L. Svennerholm).
Advances in Experimental Medicine and Biolog~ I25:33 tplenum Press) New York.

~22-o~ `
~ o . . LJ
.~ ~ _ ~J
V~ ~ . tD ' O ~ O' O O O ~r~ o ~C~ oooo a~o ~ ~ ~ 1 ~ 3 0 ¢
~ ~
U~
t~ ~1 ~
l U~ ~ ~ ~
~ ~1! ~ . tJ r~ ' C~ ~
1 E t`O 3 ~ ~ -' Cl U~
~IJ V~ 1_1 3 .-1 h 11~ -g ~ E~ _ _~
O ~ S~ ~ O O O O O 3 r~ V
`.~ ~ t:4 ~ I O O
~1~? ~ O) V V r-l ~ [~
C1'11 . ~ 5~ ~ . t~
o E ~ W . U E
u~~d W . ~J
C~
- _ E
Ul ~
J ~J ~ . o 3 C
V I ~ ~
1.~ ~ C~ C
O O O O O ~0 X ~ O O O O ~J ~
t~O I ~ ~1 _1 ~1~1 ~ ~ '~I
I ~C .~ ~
~ t~
~: . . ~1 C Cl Ll . ._~ U
~-~ '~O

3 C~
~U ~J O o ~O ~ _ . . ~ CL C
a ~1 ~ t.~ ~1 0 u~
" ~ ~ s: c) ~1 ,a, ~ ,~
E ~rl to '~ ' ~rl Cl CD ~J
. S~ ' ~ O C~ U IJ ~I
a ~ s~ ~ _l ,1 ~rl c ~ ~
~d t~ ~ ~ ~3 tO t~
o) ~ ~ c c a~to ~:
C ~1 ~ ~ J,J~J ~ ~J ~
o tn tn tn tn :1 U r C
o o c LJ ~,~
- ~ r u u .
v o ~ ~
Q~ C ~ ~ u u~ o ~
u~ ~ 11 eJ
~¢
~
o t~ ~ ~ u .. cn co ~D ~ Co ~ V
t~ . ": .D L~ .
o ~ l~ r;
_._ . ~ O U~
. ~ ~ a o u ~
~: ~ t~ z ~ ~r~ ~ t!t ~ _ ., .

~2~

_ TARLE II. Reactivity of AbR24 with gan~liosides isolated from various cell lines and ti-ssues as de~ermined by four serological test systems Source of 12 gangliosides Passive hemaggln. 2 Inhib 3 G~IIA4 5I-PA TLC
r~e Zanoma (surgica Z
specimens ):
SIl ' + + +
. ~EL-JI + + + +
~L-LO + +
~feZa710ma CeZi Lines:
SK-~EL-13 + +
SK-MEL-21 ~ +
SK-MEL-28 + + + +
SK-P~L-31 ~ +
SK-MEL-37 . + + +
SK-?~EL-64 +
SK-~EL~93 + +
~leWo ~ ~ + +
Carci7toma Ce Z Z Line~:

SK-RC-7 ~ ~ ~ +
SK-RC-ll 3L.~DDER
~33J
T-2~ _ _ +
RT 4 ~ ~ ~

.. ...
-~4-LUNG
SK-LC-LL ~ - +
CERVIX
~IE 180 . _ _ +
COLON
HT-~.9 Other CeZZs a1~d ~issues.
- ASTROCYTO~ CELL LINES
AJ +
AS ~- ~
I~OLT-4 (leukemia cell llne) + + +
RAJI (lymphoma cell line) - +
BRAIN
Bovine - - - +
~louse - - - +
~ish Human (ad~i~) - - - +
Human (fetal 10 wks) +
Human (fetal 12 & 2~ wks) - - +
CHOROID (bovine) - + +
EYE
~louse Flsh LIVER
~louse lluman (fetal) - - -Human ~adult) ....

;. ~2~
.-- .

SPL~N
~louse - -Human (fetal). - -Human (adult) MUSCLE (fetal human) - ~ +
KIDNEY
llouse Human (adult) - - +
HEART ~mouse) - - - +
THY~US (mouse) LUNG
~louse ~ ~ ~
Human (fetalor adult) - ~ + +
U~IBILICUS ' +
ERYIHROCYTES
Human (A ~ B) +
Human (O) Horse Sheep Cat YLAC~NTA thuman) - - - +
G~g 1iosides:
R~b-re3ctive ~lycollpid + + +

~1 _ ~2 . ..
_ -~6-Dla Tl ICells and tissues are human in origin unless indicated.
2Passive hemagglutination with glycolipid-coated RBC. AbR24 was used at a dilut:ion of 1:100; a minumum of 5 ~g of GD3 could be detected.
3Inhibition of PA-MHA reactivity of R24 antibody (1:80,000) with SK-~EL-'8 target cells. Results were scored positive (~) when the degree of rosettin~
was less than ~0%. At this dilution, Ab~24 could be completely inhibited by ~ ~g of GD3.
4Glycolipid-mediated immune adherence (GlllA) assay. A reaction was considered positive when greater than 50~ of the surace area o the well was covered by a lawn of pro`tein A-con~ugated erythrocytes. Abl~24 was used at a dilution o 1:1000. With this amount of antibody ~ 25 ng of GD3 could be detected.

125I~PA-TLC~ In this procedure 6 ~g of lipid-bound NANA was separated by TLC and the plate treated with AbR24 (1:1500) and 1~51-protein A. ~eactive co~,ponents were detected by autoradiog~aphy. Thls procedure can detect approximately 10-25 ng of GD3-. _ _ -~7-.

Claims (17)

The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:-

1. Method of determining the presence of GD3 disialoganglioside in a test sample comprising reacting antibody R24 (AbR24) with the test sample and determining the extent of the reaction.

2. Method of claim 1 wherein a passive hemag-glutination assay is used, the reacting step compris-ing mixing a suspension of the test sample with AbR24, the extent of reaction being determined by agglutination scored visually.

3. Method of claim 1 wherein an antibody inhibition assay is used.

4. Method of claim 1 wherein a gylycolipid-mediated immune adherence assay is used.

5. Method of claim 1 wherein a method compris-ing the detection of seroligically reactive glycolipid after separation by thin layer chromatography, is used.

6. Method for recognition of GD3 disialogan-glioside in or on a human specimen which comprises contacting a GD3 specific monoclonal antibody derived by immunization with SK-MEL-28 (deposited as ATCC
HTB 72) human melanoma immunogen with a human test specimen and assaying for the presence or absence of an immune complex formation between said GD3-specific antibody and the human test specimen.

7. Method of claim 6 wherein the human test specimen is selected from the group consisting of melanoma, melanocyte, brain and astrocytoma specimens.

8. Method of claim 6 wherein the presence or absence of the immune complex formation is assayed using an assay selected from the group consisting of 125I-protein A-thin layer chromatography, glyco-lipid-mediated immune adherence, inhibition of roset-ting and passive hemagglutination assays.

9. Method of claim 6 wherein the human test specimen is a malignant melanoma and the presence or absence of the immune complex formation is assayed for using passive hemagglutination or inhibition of rosetting.

10. Method of claim 6 in which the GD3 mono-clonal antibody is derived from the human melanoma cell line immunogen SK-MEL-28 by forming hybridoma cells between mouse myeloma cells and spleen cells from animals immunized with SK-MEL-28.

11. Method for the recognition of GD3 disialo-ganglioside in a human test specimen having elevated ganglioside or elevated GD3 glycolipid levels which comprises reacting a GD3 specific monoclonal antibody derived by immunization with SK-MEL-28 (deposited as ATCC HTB 72) human melanoma immunogen with a human test specimen selected from the group consisting of melanoma, melanocyte, brain, and astrocytoma specimens and observing the presence or absence of an immune complex formation.

12. Method for determination of the presence of GD3 disialoganglioside in human melanoma or melano-cyte cells in a human test specimen which comprises reacting a GD3 specific monoclonal antibody derived by immunization with SK-MEL-28 (deposited as ATCC
HTB 72) human melanoma immunogen with the test speci-men and observing the presence or absence of an immune complex formation.

13. Method for the diagnosis of melanoma characterized by the presence of GD3 disialoganglio-side in a human test specimen which comprises reacting a GD3 specific monoclonal antibody derived by immuni-zation with SK-MEL-28 (deposited as ATCC HTB 72) human melanoma immunogen with a melanoma test specimen and observing the presence or absence of an immune complex formation.

14. Method for the diagnosis of disorders characterized by elevated ganglioside levels in humans which comprises contacting a human test speci-men with a suspected elevated ganglioside level with a GD3 specific monoclonal antibody derived by immunization with SK-MEL-28 (deposited as ATCC
HTB 72) human melanoma immunogen and observing the presence or absence of an immune complex formation.

15. Method of claim 14 wherein the elevated ganglioside level is due to an elevated disialogan-glioside level.

16. Method of claim 15 wherein the elevated disialoganglioside level is due to an elevated GD3 level.

17. Method of claim 14 wherein the disorder is malignant melanoma.