A monoclonal antibody with reactivity to asialo GM1 and murine natural killer cells?

Molecular Imnwnology, Vol. 24, No. I. pp. 57-63. 1987 Printed in Great Britain.

Ol61-589Oj87 $3.00 + 0.00 Pergamon Journals Ltd

A MONOCLONAL ANTIBODY WITH REACTIVITY TO ASIALO GM, AND MURINE NATURAL

KILLER CELLS*

FLORA R. SOLOMON and TERRY J. HIGGKNS~

Department of Microbiology, University of Pennsylvania, School of Medicine, PA 19104, U.S.A.

(Received 21 May 1986: accepted 16 June 1986)

Abstract-A monoclonal antibody (MAb) was prepared by the fusion of murine SP,-0 myeloma cells with BALB/cByJ spleen cells that were immunized with the glycolipid asialo GM, adsorbed to naked Salmon&. The specificity of the IgM antibody obtained was defined using various glycolipids, cell extracts and saccharides in ELISA assays and thin-layer chromatography (TLC) imm~oblots. The non-reducing terminal galactose is the immunodominant residue for this antibody; however, there is undetectable reactivity to free galactose, galactosylceramide or compounds with an a-linked galactose. The SH-34 antibody specifically lyses asialo GM,-expressing macrophages in the presence of complement and removes NK cells in vitro from spleen cell populations. When the specificity of the MAb was compared to that of a commercially available rabbit antiserum to asialo GM,, it was found that both cross-reacted with GM, and asialo GM, at high antibody concns; however, the MAb did not bind asialo GM, while the rabbit antiserum showed substantial reactivity to this giycolipid. It is anticipated that this MAb will be useful for the study of murine and rat natural killer cells.

INTRODUCTION

Natural killer (NK)j cells are able to spontaneously lyse several types of tumor cells without requiring prior immunization. Due to their possible importance for in zlirjo tumor surveillance, NK cells have received considerable attention in mouse, man and rat model systems (Ortaldo and Herberman, 1984; Hoshino et al., 1984). It was discovered several years ago that the glycolipid asialo GM, could be used as a marker for murine and rat NK cells (Kasai et al., 1980; Young et al., 1980; Reynolds et al., 1981) and a rabbit antiserum to asialo GM, is currently used to identify or remove NK cells in vizw or in vitro. While polyclonal antisera continue to be useful reagents, hybridoma technology has provided a more refined and consistent source of antibody reagents. We report here on the generation and characterization of an IgM monoclonal antibody which is strongly reactive to the glycolipid asialo GM,. This antibody, however, does have measurable cross-reactivity to asialo GIM, and paragloboside and to a minor degree with the parent ganglioside GM,. When the commercially available rabbit antiserum was tested using the same

*Supported in part by National Science Foundation Grant No. PCM-841 I1 10 and B.R.S.G. S07-RR-0541524 awarded by the Biomedical Research Support Grant Program, Division of Research Resources, N.I.H.

*Author to whom correspondence should be addressed. iAbbreviations: NK, natural killer: TLC, thin-laver chro-

matography: FCS, fetal calf serum; PBS, phosphate- buffered saline; C, chloroform; M, methanol; W, water; BSA, bovine serum albumin; HBSS, Hank’s balanced sa!t solution; C’, complement; MAh. monoclonal antibody; r.t., room temp.

assay system, substantial activity to asialo GM,, asialo GM, and GM, was also observed at lower dilutions; therefore, the MAb reagent compares favorably to the commercial reagent in specificity. The MAb is capable of lysing activated macrophages which express asialo GM, (Mercurio et al., 1984) and of removing NK activity from murine spleen cell suspensions in vitro. It is anticipated that the MAb to asialo GM, will provide a more precise and useful reagent for the study of murine and rat NK cells.

MATERIALS AND METHODS

Mice

BALB/cByJ and CRA/CaJ mice were obtained from the Jackson Laboratory, Bar Harbor, ME, and maintained at the Children’s Hospital of Philadelphia Animal Facility. Mice of both sexes were used at 6-10 weeks of age.

Antisera

Affinity purified peroxidase-conjugated goat antisera against mouse IgM and IgG + M were purchased from Tago, Inc. (Burlingame, CA) and peroxidase-conjugated anti-mouse IgG from Zymed Labs (San Francisco, CA). Peroxidase-conjugated anti-rabbit IgG and purified rabbit antisera against mouse kappa or lambda light chains were obtained from Boehringer-Mannheim, Indianapolis, IN. Io- dinated F(ab-), sheep anti-mouse Ig was obtained from Amersham (IM.1310; Arlington Heights, IL) and the rabbit anti-asialo GM, serum was purchased from Wako Chem. Co. {Dallas, TX).

57

58 FLORA R. SOLOMON and TERRY J. HIGGINS

Gljrolipids and carbohydrates The saccharides used in the inhibition assay were

obtained from Sigma Chemicaf Co., St. Louis, MO. with the exception of D-gafactose (~afbiochem- Behring) and D-lactose (J. T. Baker).

I13NeuAc-GgOse,cer (GM,), IV’NeuAc, I13NeuAc- GgOse,cer (GD,,), GbOse,cer (ceramide trihexoside, CTH), and GbOse,cer (globoside) were purchased from Supelco, Inc. (Bellefonte, PA). Lactoneotetra- glycosylceramide (paragloboside) was a generous gift from Dr R. Yu (Yale University). GgOse,cer (asiafo GM,) and GgOse+er (asiafo GM,) was prepared as previously described (Higgins, 198.5).

GgOse,cer (asialo GM3) was generated from asialo GM, using fi-N-acetylhexosaminidase (Sigma, Cat. No. A2264). One milligram of asialo GM, in 2 ml of 50 mM citrate buffer, pH 4 containing 1 mg/mI tauro- deoxycholate (Sigma) was incubated with 20 units of the glycosidase after first dialyzing the enzyme against citrate buffer to remove excess ammonium sulfate. The glycolipid-enzyme mixture was slowly rotated for 44 hr at 37“C, then lyophilized and dis- solved in 2ml of C/M 2:l (v/v). GlycoIipids were partitioned with 0.4ml of water, the lower layer washed once with theoretical upper phase and the upper phase discarded. The lower phase was dried and the detergent removed on DEAE-Sepharose ace- tate as described by Gruner et al. (1981). Final pu~fication of asiafo GM, was performed on silica gel G preparative TLC plates (Analtech, Newark, DE) using C/M/W 60: 25:4 (v/v/v) as the solvent. Asialo GM, and glycolipid standards were detected under U.V. light with 1 ,&diphenyl-hexatriene (DPH) in petroleum ether (10 mg/lOO ml), then the DPH was removed by redeveloping the plate nine times in petroleum ether (Hyslop and York, 1980). The band containing asialo GM, was scraped from the plate into a glass tube and eluted from the silica gel by adding 2 ml of C/M/W 10:10:3 (v/v/v) and soni- eating for 1 min in a Branson (Cu-6) bath sonicator. The silica was pelleted by centrifugation and the extraction repeated twice more before pooling the extracts and drying. Silica gel contaminants were removed by binding the glycolipid to a reverse phase (C,,) Prep Sep cartridge (Fisher Sci. Co.) in 4ml saline and washing with 4 ml water before eluting the asialo GM, with 4 ml of C/M 2: 1 (v/v) and 2 ml C/M f:2 (v/v) (Kubo and Hoshi, 1985). The organic fractions were pooled, dried, and assayed for sphingosine (Higgins, 1984).

The C6 hydroxyl of the terminal galactose of asialo GM, was oxidized to an aldehyde by galactose oxidase as described by Singh and Kanter (1982). Asialo GM, and galactose oxidase (Worthington Diagnostics, Freehold, NJ) were incubated with constant rotation for 9 hr at room temp (r.t.), dried and hydrophilic contaminants removed using a C,, Prep Sep cartridge. Oxidized asialo GM, was purified by preparative TLC, as above, using the solvent C/M/W 60 : 40 : 9 (v~v~v).

Structural analogues of asialo GM, were prepared after first removing the amide linked acetyl group of the GalNAc residue using the method of Higashi and Basu (1982). The de-~-acetyIated asialo GM, was then dried and purified by preparative TLC as above and re-N-acylated with either acetic anhydride (MCB, Cincinnati, OH) or butyric anhydride (Sigma), repurified by TLC, and quantitated using the sphingosine assay.

Preparation of glycoIip~d extracts

Single cell suspension of spleen and thymus from BALB/cByJ and CBA/CaJ mice were generated by homogenization of the organs in 2% fetal calf serum (FCS) in phosphate-buffered saline (PBS) with a loose fitting Ten-Broeck homogenizer then allowing the mixture to stand undistur~d for 5 min to allow tissue fragments to settle before decanting the cell mixture. The cell suspension was pelleted then the erythrocytes removed by Tris-ammonium chloride lysis and the white cells washed once with phys- iological saline before extraction. Peritoneal exudate cells were obtained as described below. Cell pellets were extracted twice, first using the method described by Gruner et al. (1981) then using Svennerholm and Fredman’s protocol (1980). The two extracts were combined, dried under reduced pressure and base hydrolyzed in 0.1 N methanolic NaOH at 37’C for 3 hr. The mixture was diluted with three parts de- ionized water, desalted using a C,, cartridge, as before, and the dry residue resuspended in 6% methanol in chloroform. Glycolipids and gangliosides were further purified by applying this mixture to a “Pasteur pipette” Bio-Sil A column then washing with 10 column vols of the loading solvent followed by elution of the glycolipids and gangliosides with 10 column vols each of C/M : 2/ 1, C/M/W: 60/30/4.5 and C/M/W: 30/60/8. All three eluates were combined, evaporated to dryness then resuspended in C/M :2/l and assayed for sphingosine.

Prepurat~on of hybrid cell tine

BALB/cByJ mice were immunized with asialo GM, adsorbed to naked Sulmonella minnesota as previously described (Higgins, 1985). Mice received i.v. injections of 0.1 nmole asialo GM, adsorbed to 40 p g SaI~~~e~fa in 0.2 ml saline on days 0, 3, 7, I 1, 15, 19 and 22, then after a rest of 185 days, mice received a booster injection and spleens were harvested l-4 days later. Spleen cells were fused to the SP,-0 myeloma cell line using Kennett’s protocol (1980) and plated in Iscoves medium (Gibco Labs, Grand Island, NY) supplemented with 15% fetal bovine serum, I5 mg/ml oxalacetic acid, 5 mg/ml pyruvic acid, 20 i.u./ml insulin, 30 mg/ml glutamine, 1.4 mg/ml hypoxanthine, 0.4 mg/ml thymidine, 4 x lo-‘A4 aminopterin. 0.05 mg/ml gentamicin, and 0.25 mg/ml fungizone. Supernatants were screened by ELISA and those hybrids producing antibody reacting to asialo GM j were cloned by the limiting dilution

Monoclonal antibody to aGM, and NK cells 59

method using 2 x lo4 unprimed BALB/cByJ peritoneal feeder cells per well. After 10 days, wells containing only single colonies were re-screened by ELISA on asialo GM, and strongly positive clones were expanded without feeder cells and further char- acterized.

Antibody quantitation

The amount of antibody contained in ascites or spent culture fluid was determined using Mancini single radial immunodiffusion. A purified monoclonal IgM antibody kindly provided by Dr F. Karush was used to construct the standard curve.

ELISA assay

Hybrids producing anti-asialo GM, were screened by ELISA using a modification of the method described by Smolarsky (1980). Asialo GM, (0.4 nmole in 24 ~1 methanol) was allowed to dry in Microtest III wells (Falcon, 3912) before blocking the wells with 1% bovine serum albumin (BSA) or gelatin in PBS for 1 hr at r.t. Dilutions of culture supernatants or ascites (50 11) were incubated in glycolipid-adsorbed wells for 2 hr before removing unbound antibody by washing four times with 0.2% BSA in PBS. Peroxidase-conjugated anti-mouse IgG + M (100 ~1) was then added to the wells and left for 1 hr at r.t. Unbound second antibody was removed with four additional washes before adding 100~1 of substrate solution containing o-phenylenediamine (I mg/ml, Sigma) and H,Oz (0.012%, Baker) in 0.1 M citrate-saline, pH 4.5. Absorbance at 450 nm was determined using an automatic ELISA reader (Ti- tertek, Flow Labs).

TLC immunoblots

Purified glycolipids and cell extracts were chromatographed on E. Merck high-performance TLC

plates (5 x 1Ocm) obtained from Pierce Chem. Co. (Rockford, IL) using the solvent C: M: W/60:40:9 containing 0.02% CaCl,, air dried, then treated with polyisobutylmethacrylate as described by Saito et al. (1985). Plates to be immunoblotted were immersed in 2% BSA in PBS for 45 min followed by a second immersion in 0.3% BSA/O.OS% Tween-20 in PBS for 30min. Blocked plates were submerged in SH-34 ascites or culture supernatant diluted in 0.5% BSA in PBS and left at r.t. for 4-6 hr before washing the plate four times (1 min each) by immersion in 0.5% BSA in PBS followed by two washes with the BSA-Tween-20 mixture. Immunoblots were then incubated with 4 x lO’cpm/ml of iodinated second antibody diluted in 0.5% BSA in PBS in a parafilm “bag” for 16 hr at 6°C washed as before, dried in air for 30 min and then in a vacuum dessicator for several hours before exposing to Kodak XAR-5 film at -70°C.

Determination of antibody speciJicity

The glycolipids listed in Table 1 were used to determine the specificity of the MAb produced. Equi- molar amounts (0.4 nmole) of each glycolipid were used in the test plate for a direct comparison of relative reactivities.

%or saccharide inhibition experiments, dilutions of each saccharide in 1% BSA in PBS were incubated with a constant amount of the MAb (l/300 dilution of ascites) for 18 hr at 4°C in microtiter wells pre- blocked with 1% BSA in PBS. Control wells contained the MAb with dilutions of asialo GM, either in micelles (bath sonicated for 2 min) or in liposomes (bath sonicated with cholesterol/lecithin/dicetyl phosphate,‘asialo GM, at a mole ratio of 2:0.5:0.1: 1). Residual antibody to asialo GM, was measured by ELISA after transferring the incubation mixtures to blocked wells in which asialo GM, had previously been adsorbed.

Table I. Glycolipid structures”

Galactosyl ceramide Glucosyl ceramide Asialo GM, Asialo GM, Asialo GM,

GM,

C&lb/I I + lceramide Gl@ I - lceramide Galp I-t 4Glcp I + lceramide GaINA@ I-+ 4Gal/I I -t 4Glcfi I -t I ceramide Gal/l I -+ 3GalNAcfi 1-t 4Galfl I -* 4Glcb I + lceramide Gal0 I -+ 3GalNAcfl I -t 4Galfi I -t 4Glcfl I + lceramide

3

t NeuNAcu 2

GD,,

Globoside

Ceramide Trihexoside

(CTW

Paragloboside

“Svennerholm nomenclature

Galj I -3GalNAcfl I +4GalP I -+4Gk$ I + I ceramide 3 3

t t NeuKAca2 NeuNAca 2

GalNAcfl I +3Gala I +4Galb I -+4Glcp I + I ceramide

Gala I +4Gal/l I -+4Glcb 1-t I Ceramide

Galfi I +4GlcNAcb I -+3Galp I -+4Glc/I I-t I Ceramide

“C&l, galactose; Glc, glucose; GalNAc, N-acetylgalactosamine; NeuNAc, N-acetyl-ncuraminic acid (sialic acid); GlcNAc, N-acetylglucosamine.


Complement lysis

BALB/cByJ mice were injected with 2ml of 4% Thioglycollate Medium (BBL, Cockeysville, MD) 4 days prior to harvesting. Peritoneal exudate cells were lavaged from the peritoneal cavity using cold PBS, washed and resuspended in Hanks balanced salt solution (HBSS, Flow) containing 25 mM Hepes and 0.3% BSA (C’ medium) at a final concn of 5 x 1 O6 cells/ml. Dilutions of ascites were mixed with 25 ~1 of cells (uninduced resident cells or thioglycollate-induced cells) and incubated for 1.5 hr on ice in round-bottom polystyrene microtiter plates (Flow). Cells were washed twice with cold medium by centrifugation (800 g, 1 min, 4°C) before adding 50 ~1 of rabbit complement (Cedarlane, Low-Tox-M; Hornby, Ontario) at a l/7 dilution in C’ medium and incubating for 45min at 37°C. Cells were washed once by centrifugation with cold C’ medium then resuspended in 25 ~1 of fresh medium. Lysis was determined microscopically using a hemocytometer following the addition of 50 ~1 of 0.1% trypan blue. Per cent specific lysis was determined by the formula:

(Ab + C’) - C’ alone x 1oo

100-c’ alone ’

Natural killer cell assay

Male BALB/cByJ mice 68 weeks of age were injected i.p. with 75 pg poly 1:C (P-L Biochemicals, Milwaukee, WI) in HBSS 1618 hr before harvesting spleen cells as described above. Spleen cell suspensions were incubated with SH-34 ascites or control ascites and complement as described above before testing for the presence of NK cells using YAC-1 target cells in a 6 hr “Cr release assay. Results are expressed as specific lysis obtained from the formula:

experimental - spontaneous release

detergent lysis - spontaneous release x 100.

asialo GM, M

GM, M

GDI, --x

--*--? 100 300

IK sK ‘Koz” ‘P” ASCITES FLUID DILUTION -1

Table 2. Characteristics of the anti-asialo GM, monocloncal antibody (clone SH-34)

Heavy chain Light chain Ab titer (ascites) Ab concn:

cell culture fluid ascites fluid

P

1,30,00~,, 100,000

55 pglml I.8 mg/ml

RESULTS

Production of anti-asialo GM, hybridoma

Mice were immunized via the lateral tail veins on days 0, 3, 7, 11, 15, 19, 22 with 0.1 nmole asialo GM, adsorbed to naked Salmonella, boosted on day 201 and the spleen cells were removed and fused with Sp,-0 myeloma cells 3 days later. The fusion yielded 37 colonies of which one showed substantial anti- asialo GM, activity. The culture supernatants and later the ascites fluids of the cloned, positive colony (SH-34) were analyzed by ELISA and found to contain an IgM antibody with kappa light chains. Antibody concns of 55 pg/ml in culture supernatants and 1.8 mg/ml in ascites fluids were determined using the Mancini method. Titration of the antibody in asialo GM,-adsorbed microtiter wells gave end point titers between l/30,000 and l/100,000 with ascites fluid (Table 2).

Direct binding of glycolipids

The specificity of the SH-34 antibody was tested by ELISA using several purified glycolipids. Only glycolipids with a p-linked terminal galactose were bound by the antibody, identifying galactose as the immunodominant residue [Figs l(A) and (B), 2(B)

and 3(B)]. The absence of GD,, binding indicated the importance of the terminal galactose residue and/or the involvement of the upper plane of this residue since one side of this galactose is blocked in GD,, by sialic acid at the C3 position. Reduced binding to GM, supports this interpretation since sialic acid

100 300 IK 3K I$ 3$ ‘$0

ASCITES FLUID DILUTION -1

Fig. 1. Specificity of SH-34 by ELISA Fquimolar amounts of glycolipids (0.4 nmole) were dri microtiter wells and excess sites were blocked with BSA before reacting with antibody. Afte antibody was washed away and peroxidase labeled second antibody was added for an add;

r.t. before washing and adding substrate.

Monoclonai antibody to aGM, and NK cells 61

Fig. 2. Specificity testing of SET-34 using immunobIo~. Glycolipids were chromatographed on hpTLC plates in C/M/W :~/~/9 containing 0.02% CaCI,. (A) For each gtycolipid 8 /lg was chromatographed then visualized with Orcinol reagent. (B) For each glycohpid 4pg was chromatographed before immunoblotting as described in Methods and Materials using a l/60 dilution of SH-34 ascites in 1% BSA. Following development with iodinated second antibody and a complete drying, the plate was exposed to

X-ray film for 8 hr at -70°C.

bound to the C3 position of the intemaf galacose would also sterically hinder the upper plane. No binding was observed with galactosyl- or glucosyl- ceramide but detectable binding of asialo GM3 was observed, indicating the requirement for additional saccharide residues either as “spacers” for the galactose residue or as specific components for recognition by the MAb. The lack of binding to CTH further supports the hypothesis that the galactose must be j-linked for recognition by this antibody.

band, which co-migrates with authentic asialo GM,, was bound by the MAb when the cell extracts were probed with SH-34 [Fig. 3(b)].

The SH-34 antibody was also tested for its ability to agglutinate erythrocytes from several sources. No agglutination of sheep, guinea-pig or human A, B or 0 red cells was observed (data not shown) su~esting that the antibody does not recognize Forssman glycolipid (sheep rbc), asialo GM, (guinea-pig rbc) or any of the human blood group glycolipids.

The selectivity and specificity of the MAb was Studies involving structural analogues of asialo further tested using TLC immunoblots of whole cell GM, (Figs 4 and 5) demonstrated two points. First, extracts as well as purified glycolipids. Figure Z(B) the lower plane of the ~nultimate sugar was not shows that the same pattern of reactivity was ob- involved in antibody binding since substitution of a served using this alternate technique as seen with the butyryl group for the acetyl group normally bound to ELISA assay and Fig. 3(B) shows that the glycolipid nitrogen at the C2 position of N-acetyl- paragloboside which has a galactose ~-linked to an galactosamine (GalNac) did not alter antibody bind- N-acetylglueosamine (GlcNAc) residue was also ing. The titration curve of the antibody on free bound by the antibody. More importantly, the select- amino-asialo GM,, however, was enhanced, but this ivity of SH-34 for asialo GM, under conditions where may be due to a charge effect. Second, modification it will be employed, were best demonstrated using of the C6 hydroxyl of the terminal galactose immunoblots. Peritoneal exudate, spleen and thymus significantly decreased antibody binding, adding fur- cells each contain several glycolipids identifiable on ther support to the hypothesis that the terminal TLC using a carbohydrate specific stain [Fig. 3(A)]. galactose is the dominant residue and that the upper However, of the many glycolipids present, only one plane of the antigen is involved in antibody binding.


Fig. 3. Reactivity of SH-34 with cell extracts. Conditions as for Fig, 2. (A) Two micrograms each of asialo GMI and asialo GM,, 1 pg of paragloboside and 0.9-1.4nmoles sphingosine of cell extracts were chromatographed then sprayed with Orcinol reagent. (B) Two micrograms each of asialo GM, and asialo GM,, 0.5 pg paragloboside and 0.35-0.46 nmoles sphingosine of cell extracts were chromatographed and immunoblotted using SH-34 culture supernatant diluted l/2 with 0.5% BSA in PBS. After incubation with

iodinated second antibody and drying, the plate was exposed to X-ray film for 7 hr at -70°C.

STRUCTURAL ANALOGUES OF Asiato GMI

-1 Ceramide

[I] asialo GM1 4 re-N-acetylated-asialo GM1

[2] free amino-asialo GM1

[3] N-buty~lated asiolo GM1

-4Glc~l---1 Ceramide

[4] CG-aldehyde-asialo GM1

Fig. 4. Structural analogues of asialo GM,.


[II asialo GMt W

PI free amino-asialo GMt O--O

[I] re-N-ocetylated-asialo GMt x-x

[S] N-butyryloted-osiolo GMt H

[4] CG-aldehyde-osialo GMt A-A

ASCITES FLUID DILUTION-’

Fig. 5. Reactivity of SH-34 with structural analogues of asialo GM,. ELBA performed as for Fig. 1.

Saccharide inhibition studies

In order to further study the importance of the penultimate residue, several di- and oligosaccharides with terminal galactose residues were tested in an inhibition assay (Table 3). Galactose and l-methyl galactosides were not inhibitory, ruling out the possibility that the SH-34 MAb is simply an anti-galactose antibody. Inhibition of antibody binding was detected only with a greater than 300”fold excess of saccharide compounds containing galactose bound /?l-4 to glucose, mannose, or fructose. Consistent with the results obtained with CTH, compounds containing an @-Iinked galactose were even less inhibitory than those containing a P-linked galactose.

SH-34 and complement @sis of asiafo GM, bearing

C&S

The SH-34 antibody was tested for its ability to recognize and kill, by complement lysis, cells known to express asialo GM,. Thioglycollate-induced macrophages have been shown to express asiaIo GM, while uninduced macrophages do not have this structure exposed to the external milieu (Mercurio et al., 1984). When tested with SH-34, resident BALB/cByJ peritoneal cells were not affected by the MAb while up to 29% of the peritoneal exudate cells induced by thioglycollate were Iysed in the presence of the MAb and complement (Table 4). These results indicate that the SH-34 antibody is capable of specifically binding cells bearing exposed asialo GM, and is unreactive to normal peritoneal exudate cells.

The MAb was next tested for its ability to eliminate NK activity from the spleen cells of poIy I : C stimu- lated mice. The results presented in Table 5 show that the SH-34 MAb is very effective at eliminating the NK activity in mouse spleen cell populations. Fol- lowing C’ treatment the number of dead cells in the SH-34 treated groups (l-S%) did not significantly differ from that observed with complement alone

(4%) or with the control ascites (7%) indicating, as expected, that a very small, specific population of cells is eliminated by the antibody.

Comparison of poIyclonai antiserum and MAb reactivities

A commercially available rabbit antiserum to asialo GM, and the SH-34 monocloncal antibody were compared using ELISA with several test glycolipids. Anti-asialo GM, activity was the strongest component of both preparations (Fig. 4). While anti-GM,

Table 3. Inhibition of MAb-asialo GM, binding’

Inhibitor resulting in 50% inhibition

Inhibitor Inmotes)

Asialo GM, (micelies) Asialo GM, (iiposomes) o-galactose o-lactose (o-Galb I -t4a-Glc) D-Gal@ I -3wAra” o-Galp I -t4o-Man o-lactulose (o-Galj 1~4~.Fru) o-stachyose (~-Gala I ABo-Gala I -r6

~-Gala I *IFru) I-O-methyl-a-o-Gal

0.34.6 <0.3 1250

62.5-125 >250

I25 125250

> 250 > 250

250

Various amounts of each inhibitor were incubated with a l/300 dilution of W-34 ascites in PBS + I % BSA overnight at 4°C; transferred to blocked wells containing asialo GM, (0.4 nmoles) and residual antibody measured by ELISA.

“Ara. arabinose; Man, mannose; Fru, fructose: other abbreviations as for Table 1.

Table 4. Per cent specific lysis of peritoneal macrophages”

Induced cells” Resident cells’ ---___.._--

Ascites Ascites Dilution SH-34 Control SH-34 Control

I/50 27 0 I 1 I !250 26 1 0 0 l/l250 29 0 0 l/6250 I5 l/31250 7

Various dilutions of SH-34 ascites or a control ascites were incubated with 1.25 x IO5 peritoneal exudate cells (thioglycollate-induced or uninduced) before adding complement and determining lysis by trypan blue exclusion.

“Thioglycollate-induced cells contained 80% macrophages.

‘Resident cells (uninduced peritoneal cells) contained 37% macrophages.

Table 5. Removal of NK activity by SH-34 antibody plus com- niement

% YAC-I lysis

Treatment E:T”

Ascites dilution 1OO:l so:1

Complement alone - 39 32

SH-34 i- C’ I/IS 0 0 SH-34 + C If30 0 0 SH-34 + C l/60 3 0

Control ascites” C l/IS 39 29 Control ascites” + C l/30 44 38 Control ascites* + C’ I /60 35 35

4Effector (spleen cells) to target ratio. hAscites contained antibody to an irrelevant viral protein.


SH 34 - asiolo GM, A-A GM, AWI asialo GM2 w osiolo GMJ

w

:

B R a asialo GM, OFF

SCALE

I .6

1.2

::

t 0.8

0.4

50 200 800 3200 12800 51200

DILUTION-’

Fig. 6. Comparison of SH-34 MAb and polyclonal rabbit anti-asialo GM, serum specificity. ELISA performed as

described for Fig. 1.

activity was significant at high concns of both antibodies, both titrated to zero at a point where the preparations still exhibited high anti-asialo GM, activity. The SH-34 antibody showed greater anti-asialo GM3 activity than the rabbit antiserum; however, the MAb does not react with asialo GM, while the rabbit antiserum displays considerable reactivity to this glycolipid. These results indicate that the MAb and the anti-asialo GM, component of the polyclonal antiserum have similar cross-reactivities. The anti- asialo GM, activity present in the rabbit antiserum most likely represents a separate set of antibodies which should be taken into consideration if the polyclonal antiserum is used without additional purification.

DISCUSSION

A hybrid cell line producing an IgM monoclonal antibody recognizing the glycolipid asialo GM, was generated after multiple immunizations of BALB/cByJ mice with the antigen bound to naked Salmonella. Specificity studies indicated that the terminal, unsialylated galactose is the major recognition site for this antibody. Involvement of the upper plane of the antigen molecule and of the penultimate saccharide was also indicated. Using structural analogues of asialo GM,, additional evidence for the importance of the upper plane of galactose and a b-linkage was obtained. These studies also demon-

strated that the lower plane of the GalNAc is not involved in antibody binding.

A recent publication documented the presence in humans of natural anti-galactose antibodies (Galili et al.. 1984). The measurable reactivity of SH-34 to asialo GM, containing a galactose residue linked to a glucose raised the possibility that the SH-34 MAb was a murine anti-galactose antibody. The absence of reactivity to galactosyl-ceramide mitigates against this possibility; however, the lack of binding could be due to a lack of accessibility since the galactose is linked directly to the ceramide residue which is presumably intercalated into the pores of the plate. Consequently, inhibition of SH-34 in solution using galactose linked to various other saccharides was carried out to aid in defining antibody specificity. Saccharide inhibition studies on SH-34 showed that the antibody had no affinity for galactose alone and extremely low affinity to galactose linked to inap- propriate additional saccharides. thereby eliminating the possibility that this antibody would be simply anti-galactose. These results further demonstrated the importance of a p-linked galactose for reactivity.

One goal of this work was to generate a useful reagent for the study of cells expressing asialo GM, such as NK cells and activated macrophages. The SH-34 hybridoma antibody specifically lysed activated macrophages which have been shown to bear exposed asialo GM, (Mercurio et al., 1984), however, lysis of this population never exceeded 29%. Mercu- rio et al. (1984) did not determine what percentage of induced cells expressed asialo GM, nor in what density the antigen was expressed by positive cells. Therefore, the most likely interpretation of our complement lysis results is that there exists a heterogeneity of exposure and/or density of asialo GM, in these populations. Additional experiments showed that natural killer cells can also be eliminated from spleen cell populations using this monoclonal anti-

body. Comparison of SH-34 to a commercially available

polyclonal antiserum which recognizes asialo GM,

using the same test system showed that they had similar cross-reactivities to asialo GM, and GM, at high antibody concns. Cross-reactivity to asialo GM,, however, was non-existent for the MAb but was a substantial component in the polyclonal rabbit antiserum. These results shown that the MAb described here compares favorably to the rabbit antiserum in specificity, and like the rabbit antiserum,

can also be used to lyse cells expressing this glycolipid. The utility of the MAb was further demonstrated by its ability to discriminate asialo GM, in total glycolipid extracts of several mouse cell populations. These latter experiments showed that even

though SH-34 can bind glycolipids other than asialo GM,, this glycolipid is the only recognizable antigen present in mouse cell populations where the antibody will generally be used. The anti-asialo GM, hybridoma also provides a stable, consistent source of


antibodv which can be vroduced in vitro or in z+o. Kasai M.. lwamori M., Nagai Y., Okumura K. and Tada

contributing to a preference for this reagent. T. (1980) A glycolipid on the surface of mouse natural

After completion of this work. a paper appeared killer cells. Eur. J. Immun. 10, 175-180.

(Miller et al., 1986) describing a MAb raised against Kennett R. (1980) In Monoclonal Antibodies (Edited by

Kennett R.. McKearn T. and Bechtol K.), pp. 365-368. human neuraminidase-treated erythrocytes which re- Plenum Press, New York. acts to the terminal disaccharide of asialo GM, and Kubo H. and Hoshi M. (1985) Elimination of silica gel for

binds murine NK cells. No data on the fine specificity gangliosides by using a reversed-phase column after pre-

of this antibody to other glycolipids was presented; parative thin-layer chromatography. J. Lipid Res. 26,

therefore, how the MAb described by these authors 638-64 I.

Mercurio A. M.. Schwarting G. A. and Robbins P. W. compares to the one described herein cannot be (1984) Glycolipids of the mouse peritoneal macrophage.

evaluated at this time. J. exp. Med. 160, 114-1125. Miller V. E.. Lagarde A. E.. Longenecker B. M. and

Acknowledgements-The authors would like to thank Linda Higgins for excellent secretarial assistance.

REFERENCES

Galili U., Rachmilewitz E. A., Peleg A. and Flechner I. (1984) A unique natural human IgG antibody with anti- alpha-galactosyl specificity. J. exp. Med. 160, 15 19-l 53 1.

Gruner K., Van Eijk R. and Muhlradt P. (1981) Structure elucidation of marker glycolipids of alloantigen-activated murine T lymphocytes. Biochemistry 20, 45184522.

Higashi H. and Basu S. (1982) Specific 14C labeling of sialic acid and N-acetylhexosamine residues of gly- coshpingolipids after hydrazinolysis. Analyt. Biochem. 120, 159-164.

Higgins T. J. (1984) Simplified fluorometric assay for sphingosine bases. J. Lipid Res. 25, 1007-1009.

Higgins T. J. (1985) Murine antibody response to the glycolipid asialo GM, adsorbed to Salmonella. Molec. Immun. 22, 1265-1271.

Hoshino T., Koren H. S. and Uchida A. (Editors) (1984) Natural Killer Acticity and Its Regulation. Excerpta Med- ica, Amsterdam.

Hyslop P. and York D. (1980) The use of 1,6_diphenylhexatriene to detect lipids on thin-layer chro- matograms. Analyt. Biochem. 101, 75-77.

Greenberg A. I(. (1986) A phenyl$-galactoside-specific monoclonal antibody reactive with murine and rat NK cells. J. Immun. 136, 2968-2974.

Ortaldo J. R. and Herberman R. B. (1984) Heterogeneity of natural killer cells. A. Rev. Immun. 2, 359-394.

Reynolds C. W.. Sharrow S. O., Ortaldo J. R. and Herb- erman R. B. (1981) Natural killer activity in the rat. Ill. Analysis of surface antigens on LGL by flow cytometry. J. Immun. 127, 22042208.

Saito M., Kasai N. and Yu R. K. (1985) In situ immu- nological determination of basic carbohydrate structures of gangliosides on thin layer plates. Anulyt. Biochem. 148, 54-58.

Singh H. and Kanter J. (1982) Radioisotopic assay for galactocerebrosides and sulfatides. M&h. Enqm 83, 191-195.

Smolarsky M. (1980) A simple radioimmunoassay to determine binding of antibodies to lipid antigens. J. Immun. Meth. 38, 85593.

Svennerholm L. and Fredman P. (1980) A procedure for the quantitative isolation of brain gangliosides. Biochim. bio- phys. Acta 617, 97-109.

Young W. W. Jr., Hakomori S.-l., Durdik J. M. and Henney C. S. (1980) Identification of ganglio-N- tetraosylceramide as a new cell surface marker for murine natural killer (NK) cells. J. Immun. 124, 199-201.

Documents

A monoclonal antibody with reactivity to asialo GM1 and murine natural killer cells?