ARTICLE IN PRESS

Immunobiology 210 (2005) 11–21

0171-2985/$ - se

doi:10.1016/j.im

Abbreviations

HPTLC, high-p

clonal antibody�CorrespondE-mail addr

1Both author

www.elsevier.de/imbio

Generation of anti-Neu-glycolyl-ganglioside antibodies by immunization

with an anti-idiotype monoclonal antibody: A self versus non-self-matter

Ana Marıa Hernandez1, Mabel Rodrıguez1, Alejandro Lopez-Requena,Irene Beausoleil, Rolando Perez, Ana Marıa Vazquez�

Department of Antibody Engineering, Center of Molecular Immunology, P.O. Box 16040, Havana 11600, Cuba

Received 3 December 2004; accepted 14 February 2005

Abstract

We have previously generated a murine anti-idiotype (Ab2) monoclonal antibody (mAb) to a murine Ab1 mAb,named P3, which selectively binds Neu-glycolyl (NeuGc)-sialic acid on several monosialo- and disialogangliosides, andalso reacts with sulfatides and antigens expressed in human melanoma and breast tumors. This Ab2 mAb, designatedas 1E10, induced anti-anti-idiotype antibodies (Ab3) in mice and cancer patients. These Ab3 generated by 1E10 mAbwere characterized by bearing P3 mAb idiotopes (Ab3, Id+). But when the specificity of these Ab3 antibodies wastested, no specific humoral response against NeuGc-containing gangliosides was detected in sera from immunizedmice. However, hyperimmune sera from melanoma and breast cancer patients vaccinated with this Ab2 mAb were ableto react specifically with these gangliosides. The different expression of NeuGc-containing gangliosides in the normaltissues of mice and humans could explain these results. In order to demonstrate these findings in other animal specieswith a different NeuGc-sialic acid expression, we performed similar studies in monkeys and chickens. In monkeys, as inmost mammals, NeuGc-containing gangliosides are self-antigens. In contrast, chickens, like humans, lack theexpression of these antigens in normal tissues. Here we report that the antibody response against NeuGc-containinggangliosides induced by immunization with 1E10 mAb was completely different in both species. No specific antibodyresponse against these gangliosides was detected in hyperimmune monkey sera. In contrast, a strong and specific Ab3response against GM3(NeuGc) and GM2(NeuGc) gangliosides (Ab3, Ag+) was generated in chickens due to theadministration of 1E10 mAb.r 2005 Elsevier GmbH. All rights reserved.

Keywords: Ganglioside; Monkey; Chicken

e front matter r 2005 Elsevier GmbH. All rights reserved.

bio.2005.02.002

: ELISA, enzyme-linked immunosorbent assay;

erformance thin layer chromatography; mAb, mono-

; NeuGc, Neu-glycolyl

ing author. Tel.: +537 2716810; fax: +53 7 2720644.

ess: maruchi@ict.cim.sld.cu (A. Marıa Vazquez).

s contributed equally to this work.

Introduction

Gangliosides are sialic acid-containing glycolipidsthat are normal components of mammalian tissues,and have been considered attractive targets for cancerimmunotherapy based on their higher abundance intumors when compared with the corresponding normaltissues (Hakomori, 1981; Cahan et al., 1982; Hakomori,1985; Irie and Ravindranath, 1990).

ARTICLE IN PRESSA. Marıa Hernandez et al. / Immunobiology 210 (2005) 11–2112

Whereas gangliosides containing Neu-acetyl (NeuAc)-neuraminic acid are normal components of the plasmamembrane in humans, gangliosides bearing the Neu-glycolyl (NeuGc)-neuraminic acid residue are not(Higashi et al., 1977; Nishimaki et al., 1979). Thedifference between GM3(NeuGc) and GM3(NeuAc)consists only in the addition of a single oxygen atomto the NeuAc group of the sialic acid, creating a CH2OHgroup instead of CH3, in the context of a trisaccharide(Fig. 1). The absence of NeuGc-neuraminic acid inhumans is due to the inactivation of the gene forCMP-Neu5Ac hydroxylase, the enzyme responsiblefor NeuGc biosynthesis. The human protein lacks theN-terminal 104 amino acids present in the mouseenzyme due to the deletion of 92 bp in the humancDNA (Irie and Suzuki, 1998; Chou et al., 1998; Olsonand Varki, 2003). However, the presence of NeuGc-neuraminic acid residues have been reported in differenthuman tumors, detected by polyclonal and monoclonalantibodies, and also by chemical analysis (Hirabagashiet al., 1987; Higashi et al., 1988; Miyake et al., 1990;Marquina et al., 1996; Malykh et al., 2001).A recent article demonstrated the presence of small

amounts of this variant of sialic acid in some humannormal tissues and the authors claimed it was originatedfrom exogenous sources, such as foods from mammalianorigin. The higher amounts of NeuGc in carcinomaswere justified because of the higher uptake by theserapidly growing tissues (Tangvoranuntakul et al., 2003).The glycosidically bound NeuGc from exogenousglycoconjugates that enter human cells via pinocytosis,are released by the lysosomal sialidase and thentransported by the lysosomal sialic acid transporter tothe cytosol, where they are incorporated into glycocon-jugates (Bardor et al., 2005).Nevertheless, it is noteworthy that there is a

differential pattern of expression of NeuGc in normaland malignant human tissues.The murine P3 monoclonal antibody (mAb) (IgM, k),

generated by our group in Balb/c mice, reacts specifi-cally with a broad battery of NeuGc-containing gang-liosides, sulfated glycolipids and with antigens expressedin human breast and melanoma tumors (Vazquez et al.,1995a; Moreno et al., 1998; Alfonso et al., 2002).

Fig. 1. Schematic drawing of the trisaccharide structure of the

GM3 ganglioside, showing the subtle chemical difference

between the Neu-acetyl and Neu-glycolyl variants.

An anti-idiotype mAb (Ab2) was obtained byimmunizing Balb/c mice with P3 mAb (Ab1) coupledto keyhole limpet hemocyanin (KLH) in the presence ofFreund’s adjuvant. This Ab2 (IgG1, k), named 1E10,inhibited the binding of P3 mAb to GM3(NeuGc)ganglioside. It further induced anti-anti-idiotypic anti-bodies (Ab3) in syngeneic animals, but it was unable togenerate Ab3 with the specificity of P3 mAb. Theseresults suggested that 1E10 mAb was not an ‘‘internalimage’’ Ab2. In contrast, in phase I clinical trialsperformed in advanced melanoma and breast cancerpatients immunized with aluminum hydroxide-precipi-tated 1E10 mAb, the treatment was not only safe, but italso produced a specific response against NeuGc-containing gangliosides (Alfonso et al., 2002; Dıazet al., 2003).Based on these findings, we hypothesize that the

ability of 1E10 mAb to induce an antigen-specificresponse could depend on the presence or absence ofthese gangliosides in normal tissues. The inability of1E10 mAb to induce an antigen-specific response in micecould be due to the incapacity of this Ab2 to break thetolerance against these self-antigens. In this investiga-tion, we test the hypothesis using two animal models:monkeys, closer to humans in evolution and that expressNeuGc-glycoconjugates (Chou et al., 1998; Hayakawaet al., 2001) and chickens, which do not express NeuGc-glycoconjugates in their normal tissues (Fujii et al.,1982; Ledeen and Yu, 1982).

Materials and methods

Animals

Four Cynomolgus monkeys (3–4 years old) and nineLeghorn chickens (10 weeks old) were purchased fromthe Center for Laboratory Animal Production (CEN-PALAB, Havana, Cuba). Animals were housed underconventional conditions with free access to water andfood and maintained in accordance with the guidelinesstipulated by the Animal Subject Committee ReviewBoard of CENPALAB. Animal studies were performedwith approval from CENPALAB’s and CIM’s Institu-tional Animal Care and Use Committees.

mAbs

P3 mAb (IgM, k) was purified from ascitic fluid by gelfiltration chromatography using a Sephacryl S-300 high-resolution column (Amersham Pharmacia Biotech,Uppsala, Sweden) equilibrated with phosphate-bufferedsaline (PBS) containing 0.5M NaCl.1E10 mAb was purified from ascites by DEAE-exchange

chromatography followed by affinity chromatography

ARTICLE IN PRESSA. Marıa Hernandez et al. / Immunobiology 210 (2005) 11–21 13

on Protein A-CL Sepharose 4 B column and sizeexclusion chromatography using Sephadex G-25 column(Amersham Pharmacia Biotech). The specificity of thepurified antibodies was confirmed by enzyme-linkedimmunosorbent assay (ELISA) and protein concentra-tion estimated by optical density (OD) at 280 nm.The following murine mAbs were employed in

the study as controls: 8A4 mAb (IgG1, k, anti-idiotypemAb specific to an anti-GM2(NeuAc) ganglioside mAb)(Alfonso et al., 1995), ior C5 mAb (IgG1, k, anti-colorectal antigen) (Vazquez et al., 1995b) and 14F7mAb (IgG1, k, anti-GM3(NeuGc) ganglioside) (Carret al., 2000).

Gangliosides

Gangliosides, named according to the nomenclatureof Svennerholm (1964) GM3(NeuAc), GM3(NeuGc),and GM2(NeuGc) were kindly provided by Dr. L.E.Fernandez, Vaccine Department, Center of MolecularImmunology (Havana, Cuba). GM3(NeuAc) andGM3(NeuGc) gangliosides were purified from dog andhorse erythrocytes, respectively, as previously described(Stults et al., 1989). GM2(NeuGc) was purified frommouse liver (Hashimoto et al., 1983) and GM2(NeuAc)from bovine brain was purchased from Sigma (St.Louis, MO). Homogeneity and purity of gangliosideswas more than 95% as determined by high-performancethin layer chromatography (HPTLC) and densitometry(Ledeen and Yu, 1982).

mAb–biotin conjugation

Antibodies were incubated with 100 mg/ml of N-hydroxy-succinimide biotin for 4 h at room temperature.Then, 20 ml of 1M NH4Cl/250 mg of biotin was addedfor 10min to stop the reaction. Finally, mAb solutionwas extensively dialyzed with PBS (Bayer and Wilkek,1980).

Coupling of mAb to keyhole limpet hemocyanin

The coupling of mAbs to KLH was performed aspreviously reported (Raychaudhuri et al., 1986). Briefly,purified mAbs (1mg/ml) were mixed with the sameconcentration of KLH (Sigma) in PBS. Glutaraldehydesolution grade I (Sigma) was added at a final concentra-tion of 0.05%. The mixture was incubated in the darkfor 1 h at room temperature with gentle agitation, andthen dialyzed overnight against PBS.

Flow cytometry analysis

Peripheral blood mononuclear cells (PBMC) fromCynomolgus monkeys and Leghorn chickens were

isolated by Ficoll-Hypaque centrifugation. Cells,0.5� 106/tube, were incubated with 2 mg of 14F7 mAb,in PBS containing 1% bovine serum albumin (BSA), 2%autologus serum, and 0.01% azide for 30min on ice.After washing the cells were incubated with fluoresceinisothiocyanate (FITC)-conjugated goat anti-mouse im-munoglobulins (Jackson ImmunoResearch Labora-tories, West grove, PA) at 1/100 dilution for 30min onice. The cells were washed, re-suspended and analyzedby flow cytometry (Becton-Dickinson, San Jose, CA).

Induction of anti-anti-idiotype antibody (Ab3)

response

For induction of Ab3 response, two monkeys receivedfive s.c. injections at bi-weekly intervals of 2mg of 1E10mAb coupled to KLH (1E10-KLH) and emulsified forthe first dose in complete Freund’s adjuvant and inincomplete adjuvant for the other doses. Two othermonkeys were immunized in the same way, but with thecontrol 8A4 Ab2 mAb. Six chickens were immunizedwith four doses of 100 mg of 1E10 mAb emulsified inFreund’s adjuvant at 2-week intervals, while threechickens received the same treatment with the controlior C5 mAb. Serum samples were obtained before andduring immunization protocols.

Antibody binding assays

The presence of Ab3 in sera obtained from monkeysand chickens was determined in a solid-phase ELISA.Polystyrene Maxisorp microtiter plates (Nunc, Ros-kilde, Denmark) were coated with 50 ml of a solution of10 mg/ml of purified 1E10 mAb in carbonate buffer, pH9.6, and incubated overnight at 4 1C. Isotype-matchedirrelevant mAbs were used as control antibodies. Afterwashing with PBS containing 0.05% Tween 20, theplates were blocked for 1 h at room temperature withPBS containing 1% BSA. Then, diluted serum sampleswere added to each well and the plates were incubatedfor 2 h at 37 1C. After washing, alkaline phosphatase(AP)-conjugated secondary antibodies were added for1 h at 37 1C. For the detection of monkey antibodies anAP-conjugated goat anti-human IgG+IgM antibody(Jackson ImmunoResearch Laboratories) was used, asreported previously by several authors (e.g., Kang et al.,1992; Sen et al., 1997). Chicken antibodies were detectedwith an AP-conjugated rabbit anti-chicken IgY (Sigma).The plates were washed four times and the reaction wasdeveloped with a substrate solution consisting of 1mg/ml p-nitrophenylphosphate (Sigma) in diethanolaminebuffer, pH 9.8. Absorbance was measured at 405 nm inan ELISA reader (Organon Teknika, Salsburg, Austria).Binding of monkey and chicken Ab3 to purified

gangliosides was determined using an ELISA assay

ARTICLE IN PRESSA. Marıa Hernandez et al. / Immunobiology 210 (2005) 11–2114

previously described (Alfonso et al., 1995). Briefly,gangliosides (4 mg/ml) in 50 ml of methanol were driedin 96-well microtiter plates (Polysorp, Nunc). Plateswere blocked with 1% BSA in 0.05M Tris–HCl buffer,pH 7.8, for 30min at 37 1C. Serum samples wereincubated for 2 h at 37 1C, and after washing withPBS, an AP-conjugated goat anti-human IgG+IgM(Jackson ImmunoResearch Laboratories) or rabbit anti-chicken IgY (IgG) (Sigma) were added to the plates andincubated for 1 h at 37 1C. The plates were washed againand substrate solution was added. The highest serumdilution giving OD values X0.2 and being at least threetimes the value corresponding to the preimmune serumat the same dilution was considered as titer. Assays wereperformed in triplicate for each sample and thecoefficient of variation (CV) was less than 15%. TheODs of the blanks were less than 0.1.The presence of Ab3 antibodies specific to ganglio-

sides was also detected by immunostaining on HPTLCplates, as previously described (Kawashima et al., 1993).

Fig. 2. Detection of GM3(NeuGc) in PBMC of Cynomolgus monkey

GM3(NeuGc). PBMC from Cynomolgus monkeys (A and B) and

GM3(NeuGc) 14F7 mAb (A and C) or the isotype-matched control i

conjugated goat anti-mouse immunoglobulins and analyzed by flow

Briefly, HPTLC plates (Merck, Darmstadt, Germany)were used for the glycolipid fractionation. The solventsystem used for developing chromatography was chloro-form/methanol/0.2% CaCl2 in 2.5M NH3 (5:4:1, v:v:v).TLC plates were stained with orcinol reagent. Afterchromatography of 2 mg of each ganglioside, the plateswere soaked for 75 s on hexane containing 0.1% polyisobutylmethacrylate (Aldrich Chemical Company Inc.,Milwaukee, WI). After drying, the plates were incubatedwith serum dilutions for 2 h at room temperature. Afterwashing with PBS, the plates were incubated withbiotin–SP-conjugated anti-chicken IgY (IgG) or anti-human IgG+IgM (Jackson ImmunoResearch Labora-tories) for 1 h at room temperature, followed by washingwith PBS. Then, streptavidin-conjugated peroxidase(Jackson ImmunoResearch Laboratories) was addedfor 1 h at room temperature. The plates were washedagain and incubated with the substrate solution consist-ing of 40 mg/ml o-phenylendiamine (Sigma) in 80mMcitrate-phosphate buffer, pH 5.0, containing 0.12%

s and Leghorn chickens assessed by the reactivity with an anti-

Leghorn chickens (C and D) were incubated with the anti-

or C5 mAb (B and D). The reaction was developed with FITC-

cytometry.

ARTICLE IN PRESSA. Marıa Hernandez et al. / Immunobiology 210 (2005) 11–21 15

H2O2. The reaction was stopped by dipping the platesin PBS.

Antibody binding inhibition assays

In order to determine the anti-idiotype component inthe antibody response induced in monkeys and chickensimmunized with 1E10 mAb, different animal serumdilutions were incubated overnight with 500 mg/ml (finalconcentration) of isotype-matched irrelevant mAb.These mixtures were added to ELISA microplatescoated with 1E10 or the irrelevant antibody used toblock the anti-isotype response. After incubation for 2 hat 37 1C, the plates were washed and AP-conjugatedgoat anti-human IgG+IgM (Jackson ImmunoResearchLaboratories) or AP-conjugated rabbit anti-chicken IgY(Sigma) were added. The plates were incubated for anadditional hour at the same temperature, and afterwashing, the substrate solution was added. Experiments

Fig. 3. Specificity of Ab3 antibodies to 1E10 mAb in sera from immu

two monkeys (A and B, dilution 1/20,000) and two representative chi

tested against 100 ng/well of 1E10 mAb and the isotype-matched co

and chicken sera was assessed by using alkaline phosphatase-conjug

respectively. Sera were obtained 7 days after animals received the

triplicate wells. �po0:05, Mann–Whitney U-test, two-tailed.

using sera from animals immunized with the controlmAb 8A4 were performed under the same conditionsdescribed above.Inhibition of P3 mAb binding to 1E10 mAb by Ab3

sera was measured by ELISA, where different monkeyand chicken serum dilutions were incubated for 2 h at37 1C with 1E10 mAb F(ab0)2 fragments bound tomicrotiter plates. After washing the plates, biotinylatedP3 mAb at a concentration of 5 mg/ml was added andincubated 1 h at 37 1C. Plates were washed and boundP3 mAb was detected by adding avidin–biotin–perox-idase complex and developed as described above. Theadsorbance was measured at 492 nm in an ELISAreader. Percent inhibition of P3 mAb binding wascalculated relative to the binding of biotinylated P3mAb in the absence of inhibitor. Antibody titer wasconsidered as the highest dilution of sera giving morethan 20% of inhibition of the binding of 1E10 to P3mAb.

nized monkeys and chickens assessed by ELISA. Sera from the

ckens (C and D, dilution 1/40,000), immunized with 1E10, were

ntrol ior C5 mAb-coated microtiter plates. Binding of monkey

ated goat anti-human IgG+IgM and rabbit anti-chicken IgY,

fourth dose of 1E10 mAb. The data represent mean7SD of

ARTICLE IN PRESSA. Marıa Hernandez et al. / Immunobiology 210 (2005) 11–2116

The reactivity against gangliosides after pre-adsorp-tion of chicken sera with 1E10 Ab2 mAb was measuredby ELISA. Serum samples were pre-incubated with1E10 mAb overnight at 4 1C and added onto platescoated with GM3(NeuGc) ganglioside. Serum reactivitywas tested by the ELISA procedure described above.ELISA plates coated with 1E10 mAb were used as acontrol of the adsorption efficiency.

Results

Despite the several papers that report the existence ofN-glycolyl-containing ganglioside variants in the normaltissues of monkeys and the fact that the enzyme CMP-Neu5Ac hydroxylase, responsible for NeuGc biosynth-esis, appears without mutations even in all the greatapes, the primates closest to humans in evolution (Chouet al., 1998; Irie and Suzuki, 1998; Olson and Varki,2003), the presence of these antigens in Cynomolgusmonkeys, which were the ones used in our study, has notbeen reported before. For this purpose we incubatedPBMC from Cynomolgus monkeys with 14F7 mAb(Carr et al., 2000), which specifically recognizesGM3(NeuGc), and the reactivity was analyzed by flowcytometry. PBMC of Leghorn chickens were used as anegative control. As shown in Fig. 2, 14F7 mAb showeda strong reactivity against PBMC of Cynomolgus

Fig. 4. Anti-idiotype and anti-isotype response induced in monkeys

from a representative monkey (dilution 1/10,000) (A) and chicke

preincubated with the isotype-matched ior C5 mAb, and later the rea

hyperimmune serum from a representative monkey immunized with

Freund’s adjuvant, was used as control (C). �po0; 05, Mann–Whit

monkeys while no reactivity was detected againstLeghorn chicken cells. Control ior C5 mAb did notreact with any of the cells tested.We compared the specificity of the antibodies

generated in monkeys and chickens by immunizationwith 1E10 mAb. Antibody responses were tested in thesera obtained from the animals before and after theimmunization. All monkeys and chickens developed astrong humoral response against the whole 1E10 mAbmolecule. The titers of Ab3 response were higher than1/100,000, as measured by ELISA. The reactivity ofanimal sera with a murine isotype-matched mAb wasalso tested, and a higher binding to the immunizing1E10 mAb was observed in both monkey and chickensera (Fig. 3) (po0:05, Mann–Whitney U-test, two-tailed).To define the magnitude of the anti-idiotype compo-

nent in the antibody response against 1E10 mAb,hyperimmune animal sera were pre-adsorbed with anisotype-matched control mAb to block the anti-isotypecomponent in the serum. The remnant reactivity against1E10 mAb in the pre-adsorbed sera was significantlyhigher than the reactivity of the non-adsorbed seraagainst the isotype-matched control mAb in monkeysand chickens (po0:05, Mann–Whitney U-test, two-tailed) (Fig. 4A and B).To address the question if the immunodominance of

1E10 mAb idiotype could be a characteristic frequentlyshared by other IgG1 Ab2 mAbs, similar experiments

and chickens immunized with 1E10 mAb. Hyperimmune sera

n (dilution 1/40,000) (B) immunized with 1E10 mAb were

ctivity against 1E10 or ior C5 mAbs was assessed by ELISA. A

the control 8A4 Ab2 mAb, coupled to KLH and emulsified in

ney U-test, two-tailed.

ARTICLE IN PRESSA. Marıa Hernandez et al. / Immunobiology 210 (2005) 11–21 17

were performed using hyperimmune sera from monkeysimmunized with 8A4 mAb coupled to KLH. As shownin Fig. 4C, in contrast with the antibody responsebehavior detected in sera from the monkeys immunizedwith 1E10 mAb, the anti-idiotypic response generated inmonkeys immunized with 8A4 mAb was lower thanthe anti-isotypic one (po0:05, Mann–Whitney U-test,two-tailed).In addition, the presence of P3 idiotopes in Ab3

antibodies (Ab3, Id+) was demonstrated due to thecapacity of hyperimmune sera from monkeys andchickens to specifically inhibit the binding of 1E10mAb (Ab2) to P3 mAb (Ab1), as shown in Fig. 5. Noinhibition was detected when preimmune sera fromanimals were used as a control for specificity.Then, the specificity of the antibody response induced

in both animal species against gangliosides after theimmunization with 1E10 mAb was compared. Despitethe strong Ab3 Id+ response generated in monkeys dueto the immunization with the antibody coupled to KLHand in the presence of Freund’s adjuvant, no specificreactivity against NeuGc-containing gangliosides was

Fig. 5. Inhibition of binding of P3 mAb (Ab1) to 1E10 mAb (Ab2

chickens (C and D) immunized with 1E10 mAb. Binding of biotin

presence of serial dilutions of preimmune (open squares) and hyper

was calculated relative to binding of biotinylated P3 mAb to 1E10

detected (data not shown). On the other hand, specificantibody responses against GM3(NeuGc) andGM2(NeuGc) gangliosides were detected in four out ofthe six chickens immunized with 1E10 mAb in Freund’sadjuvant, even without coupling to KLH. Titers up to1/8000 were obtained. No reactivity was detected withany of the Neu-acetylated gangliosides tested (Fig. 6).The specificity of the anti-ganglioside antibody responsewas also confirmed by HPTLC immunostaining, wherean evident specific binding of hyperimmune chickenssera to the NeuGc-containing gangliosides tested wasobserved. No reaction was detected for the hyperim-mune sera from chickens immunized with ior C5 mAb(Fig. 7).Furthermore, we performed experiments where chick-

en sera were preincubated with 1E10 mAb and thenadded to ELISA plates coated with GM3(NeuGc).There were statistical differences in the reactivity againstthe ganglioside between the non-adsorbed and adsorbedsera, indicating that the immunization of the chickenswith 1E10 mAb induced antibodies characterized to beAg+Id+ (po0:05, Mann–Whitney U-test, two-tailed).

) by sera from the monkeys (A and B) and two representative

ylated P3 mAb to 1E10 mAb-coated plates was tested in the

immune (solid squares) animals sera. Percentage of inhibition

mAb in the absence of serum.

ARTICLE IN PRESSA. Marıa Hernandez et al. / Immunobiology 210 (2005) 11–2118

However, this inhibition was not higher than 40% of theoriginal binding, suggesting the presence of a fraction ofantibodies Ag+Id� (Fig. 8A). No reactivity of pre-adsorbed sera to 1E10-coated plates, used as anadsorption control, was detected (Fig. 8B).

Discussion

In this study, we compared the antibody responsesproduced by immunization with Ab2 1E10 mAb in two

Fig. 6. Specificity of Ab3 sera against GM3(NeuGc) and

GM2(NeuGc) from a representative chicken immunized with

1E10 (A) or ior C5 (B) mAbs. Reactivity of serial dilutions of

Ab3 sera was tested against different gangliosides by ELISA.

Binding was assessed by using alkaline phosphatase-conju-

gated rabbit anti-chicken IgY.

Fig. 7. HPTLC immunostaining of standard monosialogangliosid

Standard monosialogangliosides were chromatographed with chloro

visualized with orcinol staining. The same gangliosides as in A were

(dilution 1:500) from representative chickens immunized with 1E10

animal species with a differential expression of NeuGc-containing gangliosides in normal tissues (Fujii et al.,1982; Ledeen and Yu, 1982; Chou et al., 1998; Irie andSuzuki, 1998; Olson and Varki, 2003).We found that all animals of both species developed

anti-mouse immunoglobulin antibodies, but the serolo-gical response against 1E10 mAb was stronger whencompared with other isotype-matched mAb used as acontrol. The immunodominance of the idiotype deter-minants of 1E10 mAb was further demonstrated by theresults of experiments where animal sera were incubatedwith an irrelevant IgG1 mAb to adsorb the reactivityagainst the isotype. In all cases, the magnitude of theanti-idiotypic response was significantly higher than theresponse detected in non-adsorbed sera against thecontrol antibody (anti-isotypic response). This findingreinforces our previous results in phase I clinical trialswhere advanced melanoma and breast cancer patientswere treated with 1E10 mAb (Alfonso et al., 2002; Dıazet al., 2003). Although the immunodominance of someidiotypes has been reported for other Ab2 mAbs(Mittelman et al., 1990), it is not a general propertyof this kind of antibodies. In fact, here we show thatin monkeys immunized with 8A4 mAb, an IgG1Ab2 specific for an anti-GM2(NeuAc) Ab1 mAb, theantibody response against the isotype was the predomi-nant one.Besides, in both animal species, the Ab3 antibodies

generated by immunization with 1E10 mAb werecharacterized to share idiotopes with P3 mAb (Ab1),as evidenced by their capacity of inhibiting 1E10 mAbbinding to P3 mAb (Ab3 Id+). In contrast, when theantibody response against NeuGc-containing ganglio-sides induced by 1E10 mAb immunization was studied,a completely different pattern of response was detectedin monkeys and chickens. In monkeys, like in mice(Vazquez et al., 1998), no specific humoral responseagainst NeuGc-containing gangliosides was detected inthe sera of 1E10 mAb immunized animals. On the otherhand, most chickens in our study developed a specificAb3 response against GM3(NeuGc) and GM2(NeuGc)gangliosides (Ab3 Ag+), due to the immunization with1E10 mAb, as it was reported in humans (Alfonso et al.,2002; Dıaz et al., 2003). In both species, chickens and

es with Ab3 sera of representative immunized chickens. (A)

form/methanol/0.2% CaCl2 in 2.5M ammonium solution and

chromatographed as above and immunostained with Ab3 sera

mAb (B and C) or with control ior C5 mAb (D).

ARTICLE IN PRESS

Fig. 8. Reactivity against GM3(NeuGc) of chicken Ab3 sera

pre-adsorbed with 1E10 mAb. Sera from chickens immunized

with 1E10 mAb were preincubated with this antibody and the

remnant reactivity to GM3(NeuGc) (A) and to 1E10 mAb (B)

was evaluated by ELISA and compared to serum reactivity

without preincubation with 1E10 mAb. �po0; 05, Man-

n–Whitney U-test, two-tailed.

A. Marıa Hernandez et al. / Immunobiology 210 (2005) 11–21 19

humans, Neu-glycolylated gangliosides are heterophileantigens (Higashi et al., 1977; Fujii et al., 1982).In addition, like in melanoma and breast cancer

patients (Alfonso et al., 2002; Dıaz et al., 2003),Ag+Id� antibodies were produced in chickens immu-nized with 1E10 mAb. This was demonstrated by thereactivity of hyperimmune sera to GM3(NeuGc) gang-lioside after the adsorption of hyperimmune chickenssera with 1E10 mAb, which abrogate the antibodyresponse against this Ab2. The mechanism responsiblefor the generation of these antibodies is still unknown. Anatural immune network could be involved in thegeneration of this unusual antibody ‘‘parallel set’’.In the case of mice and monkeys, since NeuGc-

containing gangliosides are normal tissue components(Ledeen and Yu, 1982), tolerance mechanisms could beavoiding the immune response against these antigens.These mechanisms are still unknown, but as it has beendemonstrated for other self-antigen models (Fillatreauet al., 2002; Takahashi and Sakaguchi, 2003), onepossibility is the existence of T and/or B regulatorycells that would modulate the antibody response againstNeuGc-containing gangliosides. In this case, 1E10 mAbcould be acting as a tolerogenic ‘‘internal image’’ Ab2.Experimental evidence supporting that P3 and 1E10

mAbs are involved in a syngeneic B-T cell idiotypicnetwork has already been reported (Perez et al., 2002).In our hands, other immunogenic preparations

containing these gangliosides were able to break thetolerance against these antigens. In fact, we havegenerated not only IgM Ab1 mAbs, as P3 (Vazquezet al., 1995a), but also IgG1 mAb, as 14F7, very specificto GM3(NeuGc) ganglioside (Carr et al., 2000). Butanti-idiotype antibodies to 14F7 mAb also failed tobreak tolerance to NeuGc gangliosides in mice, whileinducing an antigen-specific-antibody response in chick-ens (Rodrıguez et al., 2003).In summary, our data suggest that these anti-idiotype

antibodies, beyond bearing antigen mimicry capacity,could behave as modulators of natural immune net-works.Experiments in mice to test the involvement of some B

and T cell populations in the regulation of the immuneresponse against self-gangliosides, and the role of P3and 1E10 mAbs in this regulation, are now in progressin our laboratory.

Acknowledgments

We thank Dr. L.E. Fernandez (Vaccine Department,Center of Molecular Immunology) for generouslyproviding gangliosides. This study was supported byRecom-Bio S.L. and by the Cuban Government.

References

Alfonso, M., Vazquez, A.M., Carr, A., Haerslev, T., Fernan-

dez, L.E., Lanio, M.E., Alvarez, C., Zeuthen, J., Perez, R.,

1995. T cell-independent B cell response to self mono-

sialogangliosides: primary response monoclonal antibodies.

Hybridoma 14, 209–216.

Alfonso, M., Dıaz, A., Hernandez, A.M., Perez, A., Rodrı-

guez, E., Bitton, R., Perez, R., Vazquez, A.M., 2002. An

anti-idiotype vaccine elicits a specific response to Neu-

glycolyl sialic acid residues of glycoconjugates in melanoma

patients. J. Immunol. 168, 2523–2529.

Bardor, M., Nguyen, D.H., Dıaz, S., Varki, A., 2005.

Mechanism of uptake and incorporation of the non-human

sialic acid N-glycolylneuraminic acid into human cells.

J. Biol. Chem. 280, 4228–4237.

Bayer, E., Wilkek, M., 1980. The use of avidin–biotin complex

as a tool in molecular biology. Methods Biochem. Anal. 26,

1–45.

Cahan, L.D., Irie, R.F., Sinyh, R., Cassidenti, A., Paulson,

J.C., 1982. Identification of human neuroectodermal tumor

antigen (OFA-I-2) as ganglioside GD2. Proc. Natl. Acad.

Sci. USA 79, 7929–7933.

Carr, A., Mullet, A., Mazorra, Z., Vazquez, A.M., Alfonso,

M., Mesa, C., Rengifo, E., Perez, R., Fernandez, L.E.,

2000. A mouse monoclonal antibody specific for

ARTICLE IN PRESSA. Marıa Hernandez et al. / Immunobiology 210 (2005) 11–2120

Neu-glycolyl GM3 ganglioside recognized breast and

melanoma tumors. Hybridoma 19, 241–247.

Chou, H., Takematsu, H., Dıaz, S., Iber, J., Nickerson, E.,

Wright, K.L., Muchmore, E., Nelson, D.L., Warren, S.T.,

Varki, A., 1998. A mutation in human CMP-sialic acid

hydroxylase occurred after the Homo-Pan divergence.

Proc. Natl. Acad. Sci. USA 95, 11751–11756.

Dıaz, A., Alfonso, M., Alonso, R., Saurez, G., Troche, M.,

Catala, M., Dıaz, R.M., Perez, R., Vazquez, A.M.,

2003. Immune responses in breast cancer patients immu-

nized with an anti-idiotype antibody mimicking Neu-

glycolyl-containing gangliosides. Clin. Immunol. 107,

80–89.

Fillatreau, S., Sweenie, C.H., McGeachy, M.J., Gray, D.,

Anderton, S.M., 2002. B cells regulate autoimmunity by

provision of IL-10. Nat. Immunol. 3, 944–950.

Fujii, Y., Higashi, H., Ikuta, K., Kato, S., Naiki, M., 1982.

Specificities of human heterophilic Hanganutziu and

Deicher (H–D) antibodies and avian antisera against

H–D antigen-active glycosphingolipids. Mol. Immunol.

19, 87–94.

Hakomori, S., 1981. Glycosphingolipids in cellular interaction,

differentiation and oncogenesis. Ann. Rev. Biochem. 50,

733–764.

Hakomori, S., 1985. Aberrant glycosylation of surface

glycoproteins in cancer cell membranes as focused on

glycolipids: overview and perspectives. Cancer Res. 45,

2405–2414.

Hashimoto, Y., Otsuka, H., Sudo, K., Susuki, K., Yamakawa,

T., 1983. Genetic regulation of GM2 expression in liver of

the mouse. J. Biochem. 93, 895–901.

Hayakawa, T., Satta, Y., Gagneaux, P., Varki, A., Takahata,

N., 2001. Alu-mediated inactivation of the human CMP-

Neu-acetylneuraminic acid hydroxylase gene. Proc. Natl.

Acad. Sci. USA 98, 11399–11404.

Higashi, H., Naike, M., Matuo, S., Okouchi, K., 1977.

Antigen of ‘‘serum sickness’’ type of heterophile antibodies

in human sera identification as gangliosides with Neu-

glycolylneuraminic acid. Biochem. Biophys. Res. Commun.

79, 388–395.

Higashi, H., Sasabe, T., Fukui, Y., Maru, M., Kato, S., 1988.

Detection of gangliosides as Neu-glycolyl neuraminic acid-

specific tumor-associated Hanganutziu–Deicher antigen in

human retinoblastoma cells. Jpn. J. Cancer (Gann) 79,

952–956.

Hirabagashi, Y., Higashi, H., Kato, S., Taniguchi, M.,

Matsumoto, M., 1987. Occurrence of tumor associated

ganglioside antigens with Hanganutziu–Deicher antigenic

activity on human melanomas. Jpn. J. Cancer Res. (Gann)

78, 614–620.

Irie, A., Suzuki, A., 1998. CMP-Neu-acetylneuraminic acid

hydroxylase is exclusively inactive in humans. Biochem.

Biophys. Res. Commun. 248, 330–333.

Irie, R.F., Ravindranath, M.H., 1990. Gangliosides as target

for monoclonal antibodies therapy of cancer. In: Borre-

baeck, C.A.K., Larrick, G.W. (Eds.), Therapeutic Mono-

clonal Antibodies. M. Stockom Press, pp. 75–94.

Kang, C., Nara, P., Chamat, S., Caralli, V., Chen, A., Nguyen,

M., Yoshiyama, H., Morrow, W., Ho, D., Kohler, H.,

1992. Anti-idiotype monoclonal antibody elicits broadly

neutralizing anti-gp120 antibodies in monkeys. Proc. Natl.

Acad. Sci. USA 89, 2546–2550.

Kawashima, I., Ozawa, H., Kotani, M., Susuki, M., Kawanu,

T., Gomobuchi, M., Tai, T., 1993. Characterization of

ganglioside expression in human melanoma cells: immuno-

logical and biochemical analysis. J. Biochem. (Tokyo) 114,

186–193.

Ledeen, R.W., Yu, R.K., 1982. Gangliosides: structure,

isolation, and analysis. Methods Enzymol. 83, 139–191.

Malykh, Y.N., Schauer, R., Shaw, L., 2001. Neu-glycolylneur-

aminic acid in human tumors. Biochimie 83, 623–634.

Marquina, G., Waki, H., Fernandez, L.E., Kon, K., Carr, A.,

Valiente, O., Perez, R., Ando, S., 1996. Gangliosides

expressed in human breast cancer. Cancer Res. 56,

5165–5171.

Mittelman, A., Chen, Z.J., Kageshita, T., Yang, H., Yamada,

M., Baskind, P., Goldberg, N., Puccio, C., Ahmed, T.,

Arlin, Z., Ferrone, S., 1990. Active specific immunotherapy

in patients with melanoma: a clinical trial with mouse anti-

idiotypic monoclonal antibodies elicited with syngeneic

anti-high molecular weight melanoma-associated antigen

monoclonal antibodies. J. Clin. Invest. 86, 2136–2144.

Miyake, M., Hashimoto, K., Ito, M., Ogawa, O., Arai, E.,

Hitomi, S., Kannagi, R., 1990. The abnormal occurrence

and the differentiation-dependent distribution of Neu-

acetyl and Neu-glycolyl species of the ganglioside GM2 in

human germ cell tumors. A study with specific monoclonal

antibodies. Cancer 65, 499–505.

Moreno, E., Lanne, B., Vazquez, A.M., Kawashima, I., Tai,

T., Fernandez, L.E., Karlsson, K-A., (Angstrom, J., Perez,

R., 1998. Delineation of the epitope recognized by an

antibody specific for Neu-glycolylneuraminic acid-contain-

ing gangliosides. Glycobiology 8, 695–705.

Nishimaki, T., Kano, K., Milgron, F., 1979. Hanganutziu–

Deicher antigen and antibody in pathologic sera and

tissues. J. Immunol. 122, 2314–2318.

Olson, M.V., Varki, A., 2003. Sequencing the chimpanzee

genome: insights into human evolution and disease. Nat.

Rev. Genet. 4, 20–28.

Perez, A., Mier, E.S., Vispo, S., Vazquez, A.M., Perez, R.,

2002. A monoclonal antibody against NeuGc-containing

gangliosides contains a regulatory idiotope involved in

the interaction with B and T cells. Mol. Immunol. 39,

103–112.

Raychaudhuri, S., Saeki, Y.S., Fuji, H., Kohler, H., 1986.

Tumor-specific idiotype vaccines. I. Generation and char-

acterization of internal image tumor antigen. J. Immunol.

137, 1743–1749.

Rodrıguez, M., Llanes, L., Perez, A., Perez, R., Vazquez,

A.M., 2003. Generation and characterization of an anti-

idiotype monoclonal antibody related to GM3(NeuGc)

ganglioside. Hybridoma 22, 307–314.

Sen, G., Chakraborty, M., Foon, K., Reisfeld, R., Bhattachar-

ya-Chatterjee, M., 1997. Preclinical evaluation in nonhu-

man primates of murine monoclonal anti-idiotype antibody

that mimics the disialoganglioside GD2. Clin. Cancer Res.

3, 1969–1976.

Stults, C.L.M., Sweeley, C.C., Matcher, B.A., 1989. Glyco-

sphingolipids: structure, biological source, and properties.

Methods Enzymol. 179, 167–214.

ARTICLE IN PRESSA. Marıa Hernandez et al. / Immunobiology 210 (2005) 11–21 21

Svennerholm, L., 1964. The gangliosides. J. Lipid Res. 5,

145–155.

Takahashi, T., Sakaguchi, S., 2003. Naturally arising

CD25+CD4+ regulatory T cells in maintaining immuno-

logic self-tolerance and preventing autoimmune disease.

Curr. Mol. Med. 3, 693–706.

Tangvoranuntakul, P., Ganneux, P., Dıaz, S., Bardor, M.,

Varki, N., Varki, A., Muchmore, E., 2003. Human uptake

and incorporation of an immunogenic nonhuman

dietary sialic acid. Proc. Natl. Acad. Sci. USA 100,

12045–12050.

Vazquez, A.M., Alfonso, M., Lanne, B., Karlsson, K.A., Carr,

A., Barroso, O., Fernandez, L.E., Rengifo, E., Lanio, M.E.,

Alvarez, C., Zeuthen, J., Perez, R., 1995a. Generation of a

murine monoclonal antibody specific for Neu-glycolylneur-

aminic acid-containing gangliosides that also recognizes

sulfated glycolipids. Hybridoma 14, 551–556.

Vazquez, A.M., Tormo, B.R., Alfonso, M., Velandia, A.,

Fernandez, L.E., Giscombe, R., Ansotegui, Y., Jeddi

Tehrani, M., Cedeno, M., Toledo, A.L., Perez, R.,

Mellstedt, H., Biberfeld, P., 1995b. Characterization of

ior C5 colorectal tumor associated antigen. Inmunologıa

14, 130–132.

Vazquez, A.M., Perez, A., Hernandez, A.M., Macıas, A.,

Alfonso, M., Bombino, G., Perez, R., 1998. Syngeneic anti-

idiotypic monoclonal antibodies to an anti-NeuGc-contain-

ing ganglioside monoclonal antibody. Hybridoma 17,

527–534.