a strategy for inducing an immune response against androctonus australis scorpion venom toxin i in...

A strategy for inducing an immune response against Androctonus

australis scorpion venom toxin I in mice. Production of

high-affinity monoclonal antibodies and their use in a sensitive

two-site immunometric assay

Christiane Devaux a,*, Olivier Clot-Faybesse a, Martine Pugniere b, Jean-Claude Mani b,Herve Rochat a, Claude Granier b

aLaboratoire de Biochimie-Ingenierie des Proteines, CNRS UMR 6560, Faculte de Medecine-Nord, Bd P. Dramard,

13916 Marseille Cedex 20, FrancebCNRS UMR 5094, Faculte de Pharmacie, Av. Charles Flahault, 34093 Montpellier Cedex 5, France

Received 4 March 2002; received in revised form 31 May 2002; accepted 13 August 2002

Abstract

Scorpion neurotoxins acting on ion channels share some structural features but differ in antigenic and immunogenic

properties. They are highly structured peptides, 60–70 amino acids long. Monoclonal antibodies have been obtained for

Androctonus australis hector scorpion venom neurotoxin II (AahII) and a nontoxic synthetic analog ((Abu)8 AahII). In this

study, no antibody response was elicited in mice of various strains injected with AahI, the other important toxin of the venom, in

a native or an inactive ((Abu)8 AahI) form. We found that AahI was only immunogenic in BALB/c or C57BL/6 mice if it was

coupled to a carrier protein. The helper protein molecule could be BSA, KLH, or the nontoxic analog of AahII. We obtained a

panel of high-affinity mAbs with these immunogens. Two of these mAbs, including the very high-affinity antibody 9C2

(KD = 0.11�10� 11 M), were used to set up a two-site ELISA, sensitive enough for the quantification of AahI in the biological

fluids of envenomed animals. The detection limit of the assay was 75 pg/ml.

D 2002 Elsevier Science B.V. All rights reserved.

Keywords: Scorpion neurotoxins; Antibody response; MHC restriction; Monoclonal antibodies; BIACORE; Immunoassays

1. Introduction

Scorpion stings cause widespread morbidity and

mortality in various parts of the world (Goyffon et al.,

1982; Amitai, 1998; Heard et al., 1999). Records

reveal an annual mean of 310 fatal accidents in Mexico

over a period from 1981 to 1992 (Dehesa-Davila and

Possani, 1994) and 13–103 cases in Tunisia from 1986

to 1997 (Krifi et al., 1999). However, the overall

number of stings was considerably higher, a fraction

of them leading to severe envenoming requiring inten-

sive care. The molecules responsible for this effect are

small basic proteins that act on sodium channel con-

ductance (Rochat et al., 1979). Immunotherapy is the

0022-1759/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved.

PII: S0022 -1759 (02 )00338 -1

* Corresponding author. Tel.: +33-4-9169-8857; fax: +33-4-

9165-7595.

E-mail address: [email protected] (C. Devaux).

www.elsevier.com/locate/jim

Journal of Immunological Methods 271 (2002) 37–46

specific medical treatment in cases of severe scorpion

envenoming. Thus, specific poly- and monoclonal

antibodies have been produced against these toxins

as a means of treating envenoming and as tools for

investigating scorpion envenoming. The venom of the

scorpion Androctonus australis hector contains three

main toxins: AahI, AahII and AahIII. Although highly

toxic, these molecules nevertheless induce a strong

immune response in rabbits, high-titer antisera being

obtained by means of a progressive immunization

schedule (Delori et al., 1981). These three toxins have

been used to develop specific radioimmunoassays

defining two immunological groups: group 1 contains

AahI and AahIII, which displays up to 80% sequence

identity, and group 2 contains AahII, which displays

only 44–45% sequence identity with AahI and AahIII.

The toxins of groups 1 and 2 display no cross-reac-

tivity with the reciprocal polyclonal antiserum (El

Ayeb et al., 1983). Monoclonal antibodies (mAb)

against the various toxins may help us to understand

the fine antigenic specificity of these molecules and the

mechanisms of neutralization. The mAbs neutralizing

AahII have been produced by immunization with

AahII (Bahraoui et al., 1988) or a nontoxic synthetic

analog (Abu)8 AahII (Devaux et al., 1997). However,

we obtained no anti-AahI mAb with these two

approaches, whatever the mouse strain used.

In this study, we describe a strategy for immuniz-

ing mice with AahI derivatives so as to obtain sera

with high immunoreactivity with native toxin both in

ELISA and RIA. A panel of mAbs was obtained and

their affinities for AahI were determined using BIA-

CORE technology. A sensitive two-site immunomet-

ric assay was designed and shown useful for

quantifying AahI in biological samples. Modifications

of the immunogen and immunization protocols are

discussed, as are the possible explanations for the

observed differences in the immunogenicity of AahI

and AahII.

2. Materials and methods

2.1. Animals and venoms

BALB/c (H-2d) and C57BL/6 (H-2b) mice were

raised and housed in the conventional animal facilities

of our laboratory. B10.M (H-2f), B10.SM (H-2v),

C3H NB (H-2p), B10.PL (H-2u), B10.S (H-2s), B10

WB (H-2j), C3H K (H-2q), B10 RIII (H-2r), and CBA

(H-2k) mice were a gift from Marika Pla (CNRS UPS

44, Orleans, France). Adult male New Zealand white

rabbits were purchased from Elevage Scientifique des

Dombes (France). Animals were cared for in accord-

ance with institutional guidelines (No. 86/609/CEE).

Venoms were obtained by manual extraction from

scorpions collected in various parts of Tunisia.

2.2. Immunogens and immunization

Toxins AahI, AahII and AahIII were obtained from

the venom of the scorpion A. australis hector. They

were carefully purified and characterized in the labo-

ratory (Miranda et al., 1970). Nontoxic analogs

(Abu)8 AahI and (Abu)8 AahII were synthesized as

previously described (Zenouaki et al., 1997). (Abu)8AahI and AahI were either polymerized using gluta-

raldehyde (Sigma, Saint Quentin Fallavier, France) or

coupled to (Abu)8 AahII, BSA or KLH (Pierce) using

glutaraldehyde or 1-ethyl3 (dimethylaminopropyl)

carbodiimide (EDC, Pierce Perbio, Bezons, France).

The molar ratio for AahI/(Abu)8 AahII and AahI/BSA

was 4:1. Groups of mice, 6- to 8-weeks old, were

injected subcutaneously (s.c.) with the appropriate

immunogen in Freund’s complete adjuvant on day 0

(D0) and then intraperitoneally with the immunogen in

Freund’s incomplete adjuvant on D15, D30, D70 and

D105. Ten days after each injection, retro-orbital blood

samples were obtained and sera were assayed by

ELISA and RIA.

2.3. Production of mAbs

Protocols for spleen cell fusion, hybridoma screen-

ing and selection, and for the mass production and

characterization of immunoglobulins were as previ-

ously described (Clot-Faybesse et al., 1999). Briefly,

splenocytes from immune mice (108 cells) were fused

with X63-F murine myeloma cells (2� 107) by poly-

ethyleneglycol 1500. Hybrids were selected either in

HAT medium, with antibody-secreting cells cloned at

limiting dilution (Fusion 1), or in a methylcellulose-

based medium (Clonacellk-HY, France; Fusion 2).

Antibodies were produced as ascitic fluid in BALB/c

mice and purified on protein A Sepharose (Amersham

Pharmacia, Orsay, France).

C. Devaux et al. / Journal of Immunological Methods 271 (2002) 37–4638

2.4. Immunoassays

The ability of each mAb to bind to AahI was

assessed by ELISA and RIA as previously described

(Devaux et al., 1997). AahI was labeled with [125I] as

previously described (Rochat et al., 1977).

2.5. BIACORE analysis of monoclonal IgG

All experiments were performed at 25 jC on a

BIACORE 2000 apparatus (Biacore, Uppsala, Swe-

den). AahI was covalently immobilized on a CM5

sensor chip via its amino groups (EDC/NHS activa-

tion) or via its carboxyl groups after derivatization with

hydrazine according to the manufacturer’s protocols.

Around 300 RU (300 pg mm2) of AahI were immobi-

lized on a flow cell. A second flow cell was treated

with buffer and activating reagents alone and used as a

control. Increasing concentrations of mAbs in HBS (10

mM HEPES buffer, pH 7.4, 3 mM EDTA, 0.005%

P20, supplemented with 150 mM NaCl) were injected

into the AahI and control flow cells at a rate of 50 Al/min, using association and dissociation phases of 180

and 400 s, respectively. The sensor surface was regen-

erated by passing 10 Al of 0.1 M HCl over the surface.

Sensorgrams were corrected by subtracting the control

flow cell signal, and the kinetic and affinity constants

were obtained from sensorgrams using BIAevaluation

3.1 software (Biacore) and the global fitting method.

Flow-independent kinetic parameters showed an

absence of mass transport and rebinding effects.

2.6. Biotin labeling of protein A-purified IgG

Monoclonal antibodies were purified from ascitic

fluids by protein A Sepharose chromatography. The

N-hydroxysuccinimide ester of biotin (NHS-Biotin,

Pierce, Rockford, IL) in DMSO was added to 1 mg of

purified IgG in 50 mM sodium carbonate buffer, pH

8.5 at a molar ratio of 80:1. The mixture was incu-

bated for 30 min at room temperature and then 2 h on

ice; it was then exhaustively dialyzed against borate

buffer saline (BBS), pH 7.9.

2.7. Two-site immunometric assay

Immunometric assays were performed in 96-well

microtiter plates (Maxisorb, NUNC). All possible

pairwise combinations of mAbs were initially tested.

The most sensitive sandwich pair was selected and the

assay was then optimized. The working concentra-

tions of the capture mAb used for microplate coating

and the biotinylated mAb used for detection were

independently chosen to reach a high sensitivity and a

low background. Diluants, times of incubation and

concentration of streptavidin peroxidase were also

alternatively assayed. A one-step procedure (simulta-

neous addition of ligand and antibody) was used. The

following conditions correspond to the optimized

assay. The wells were coated by incubation with

IgG 2G3 at a concentration of 16 Ag/ml in PBS (2 h

at 37 jC). They were then saturated by incubation

with 0.1% Tween-20, 2.5% casein in PBS (1 h at 37

jC) and washed four times. A series of dilutions of the

standard solution of AahI from 40 to 0.04 ng/ml (50

Al) were added to successive wells together with

biotin-labeled IgG 9C2 at a concentration of

5� 10� 10 M (50 Al). The plates were then incubated

for 90 min at 37 jC and washed four times. Perox-

idase-labeled streptavidin was diluted (1:800) in PBS

supplemented with 0.1% Tween, 0.1% casein and

0.5% nonimmune mouse plasma to prevent nonspe-

cific binding to heterophilic antibodies (Ho et al.,

1986) and was added to each well (100 Al). The plateswere incubated further for 45 min at 37 jC and were

then washed again. TMB (100 Al, Kirkegaard and

Perry Labs, MD) was added as a substrate and the

plates incubated for 10 min at room temperature.

Peroxidase activity was measured at 650 nm with an

iEMS reader (LabSystems). Alternatively, the reaction

was stopped by the addition of 1 M phosphoric acid

and the absorbance read at 450 nm. The assays were

carried out in triplicate.

3. Results

3.1. Analysis of the anti-AahI antibody response in

various strains of mice

The toxicity of scorpion neurotoxins limited the

amount of immunogen that could be injected to

induce an immune response. The LD50 of AahI was

0.38 Ag in 20 g C57BL/6 mice and 1.4 Ag in BALB/c

mice when injected s.c. Emulsification in Freund’s

adjuvant reduced the availability of the molecule and

C. Devaux et al. / Journal of Immunological Methods 271 (2002) 37–46 39

made it possible to inject up to 4 Ag of toxins into

C57BL/6 mice and 12 Ag into BALB/c mice without

killing the animals. In a preliminary study, we used

ELISA to monitor antibody production in the sera of

mice of strains bearing various haplotypes of MHC

class II molecules (Table 1). Of the 11 mouse strains

tested, only B10.PL (H-2u) and C57BL/6 (H-2b)

displayed a moderate response as detected by ELISA

after five injections of AahI. Several fusion experi-

ments were then performed with the spleens of immu-

nized C57BL/6 mice. These spleens were atrophied in

appearance and, after fusion, the few growing hybrids

produced nonspecific IgM.

3.2. Analysis of the anti-AahI immune response using

various designs of immunogens

We investigated whether the lack of immune

response was due to the small amounts of toxin

injected to avoid killing the animal. We prepared

and used two types of molecule for this analysis.

First, we produced, by chemical solid-phase synthesis,

a peptide containing the entire sequence of toxin I

(Abu)8 AahI with each half-cystine replaced by the

isosteric residue, a-aminobutyric acid. Second, AahI

was treated with glutaraldehyde, a reagent that has

already been used to detoxify venoms (Possani et al.,

1981; Kharrat et al., 1997). The toxicity and immu-

nogenicity of these constructs were tested in C57BL/6

(H2b) and BALB/c (H2d) mice. (Abu)8 AahI was

totally nontoxic and glutaraldehyde-polymerized AahI

preparations were detoxified by a factor of 5–10

depending on the individual preparation, as shown

by the increase in LD50. A single immunization

schedule was developed based on injections of 40

Ag/mouse. The anti-toxin antibody response was sys-

tematically evaluated at each bleeding by ELISA and

RIA against AahI. Glutaraldehyde-polymerized AahI

and (Abu)8 AahI alone gave no response or, in some

cases, a weak response in ELISA, which was not

confirmed by RIA with [125I]AahI (Table 2). There-

fore, the lack of immunogenicity of AahI is probably

Table 1

AahI-specific antibody response in 11 strains of mice differing in

MHC Class II molecules

Strain Haplotype ELISA

B10.M H-2f �B10.SM H-2v �C3H NB H-2p �B10.PL H-2u +

B10.S H-2s �B10 WB H-2j �C3H K H-2q �B10 RIII H-2r �C57BL/6 H-2b +

BALB/c H-2d �CBA H-2k �ELISA were performed after four injections of AahI in Freund’s

adjuvant. Gradation was evaluated by measuring absorbance at 450

nm for a serum dilution of 1:1000. (� ) Corresponds to A< 0.1, (+)

to A< 1 and (++) to A>1.

Table 2

Evaluation of the humoral immune response to AahI elicited in mice

by the AahI toxin or the (Abu)8 AahI analog or by modified

molecules

Immunogen Mouse

strain

ELISAa RIAb Hybridomac

AahI BALB/c � ND IgM

C57BL/6 + �Glutaraldehyde- BALB/c � ND

polymerized AahI C57BL/6 � ND

(Abu)8 AahI BALB/c � ND

C57BL/6 � ND

(Abu)8 AahI-BSA- BALB/c + �glutaraldehyde C57BL/6 � �

NMRI � ND

(Abu)8 AahI-BSA-EDC BALB/c � ND

C57BL/6 � ND

(Abu)8 AahI-KLH- BALB/c � ND

glutaraldehyde C57BL/6 � ND

(Abu)8 AahI-(Abu)8AahII-glutaraldehyde

NMRI + �

AahI-BSA-glutaraldehyde BALB/c + + + + IgG1/IgG2a

C57BL/6 + + +

AahI-(Abu)8 AahII- BALB/c + + + + IgG1/IgG2b

glutaraldehyde C57BL/6 + + +

Evaluation was carried out by ELISA and RIA.a Immunoreactivity was measured after three injections of

immunogen. Gradation was evaluated in ELISA by measuring the

absorbance at 650 nm for a serum dilution of 1:1000. (� )

Corresponds to A< 0.1, (+) to A < 1 and (++) to A>1.b Immunoreactivity was measured after three injections of

immunogen except for AahI-(Abu)8 AahII-glutaraldehyde, for

which the immune response was only detected by RIA after five

injections. Gradation was evaluated in RIA as the percentage of

[125I]-AahI bound versus total radioactivity in the assay for a serum

dilution of 1:10,000. (� ) Corresponds to B/T < 10%, (+) to B/

T < 50% and (++) to B/T>50%. ND=Not determined.c The class of immunoglobulins purified from the hybridoma is

given when obtained.


not due to the relatively low amounts used for

immunizing animals.

Finally, we tried to increase the immunogenicity of

the AahI molecule by carrying out immunizations

with (Abu)8 AahI or AahI covalently coupled to

carrier proteins. The carrier proteins used were BSA,

KLH and (Abu)8 AahII, a synthetic toxin derivative

known to be immunogenic in mice (Devaux et al.,

1997; Zenouaki et al., 1997). The coupling reagents

used were glutaraldehyde or EDC. Protein-coupled

AahI preparations were detoxified by a factor of 5–10

relatively to the free protein. The immunization pro-

tocols used were similar to those described above. If

AahI was coupled to BSA or (Abu)8 AahII, a strong

anti-AahI response was detected in both ELISA and

RIA (Fig. 1). BALB/c mice showed a stronger

response than C57BL/6 mice did. A strong signal in

ELISA did not necessarily correlate with a good

recognition of [125I]-AahI in RIA (Fig. 1). All modi-

fied immunogens based on the nontoxic analog

(Abu)8 AahI rather than native AahI gave only a faint

response, if indeed they gave a signal at all, in ELISA.

Again, this response was not observed in liquid-phase

RIA (Table 2).

3.3. Production of mAbs and real-time interaction of

mAbs with AahI measured in the BIACORE biosensor

Two fusion experiments were conducted with the

mice giving the strongest response, as assessed by

RIA, and a panel of high-affinity mAbs was produced

Fig. 1. ELISA and RIA analyses of the humoral response elicited in

mice by carrier-coupled AahI. BALB/c mice were injected with

AahI coupled to BSA (white column) or AahI coupled to (Abu)8AahII (hatched white column) and C57BL/6 mice were injected

with AahI coupled to BSA (black column) or AahI coupled to

(Abu)8 AahII (hatched black column) on days 0, 15, 30, 70 and 105.

Sera obtained 10 days after each injection were tested by indirect

ELISA (A) and by RIA (B) at a dilution of 1:1000.

Table 3

Kinetic parameters for the binding of anti-AahI mAbs to toxin AahI,

as assessed by BIACORE measurements

mAb Ka1

(M� 1 s� 1)

Kd1(s� 1) KD1

(M) Rmax

(RU)

v2

9C2 5.5� 105 5.9� 10� 6 0.11�10� 10 2400 4.3

2G3 5.3� 105 1.4� 10� 5 25.7� 10� 10 83 4.7

7D3 2.2� 105 1.4� 10� 4 6.4� 10� 10 168 4.8

10A10 1.5� 104 1.7� 10� 4 11.8� 10� 10 96 1.4

5B5 1.2� 105 6.4� 10� 4 52.9� 10� 10 101 0.2

Statistical value for describing the closeness of fit. Values of v2

below 10 are usually acceptable.

Fig. 2. Comparison of the immunoreactivities of biotin-labeled

mAbs with those of unlabeled mAbs. The reactivity with AahI in

RIA of biotin-labeled IgGs (closed symbols) from mAbs 9C2

(triangles) and 2G3 (circles) was compared with that of parent

antibodies IgG 9C2 and 2G3 (open symbols).


(Clot-Faybesse et al., 1999). According to the results

of RIA (Fig. 1B), the immunization protocols were

slightly modified. The mouse that received AahI

coupled to (Abu)8 AahII was given seven injections

of immunogen over a period of 10 months (Fusion 1,

long protocol). The other mouse was only given four

injections of AahI coupled to BSA over a period of 3

months (Fusion 2, short protocol). The best mAb

(9C2) was obtained with this protocol.

Protein A-purified IgG was prepared and their

interaction with AahI was further studied by measuring

association/dissociation rate constants in the BIA-

CORE system. AahI was covalently linked to the

dextran matrix of the sensor chip via its amino groups.

We injected mAbs at concentrations of 3.5–260 nM,

except for mAb 9C2, which was injected at concen-

trations of 0.6–6.6 nM. Table 3 gives the kinetic

constants of association (ka) and dissociation (kd) and

equilibrium dissociation constants (KD). All mAbs

gave KD in the nanomolar range, except for mAb

9C2, which had an even lower KD (0.11�10 � 11

M), calculated according to a Langmuir 1:1 model.

This was the result of a very slow rate of dissociation of

mAb 9C2 from AahI (kd = 5.9� 10 � 6 s � 1). The

association rates (ka) of all the mAbs were in the

1.2–5.5� 105 M� 1 s� 1 range in this fitting model.

Similar results were obtained if AahI was immobilized

on the sensor chip via its carboxyl groups (not shown).

The binding of mAb 9C2 was also analyzed using a

bivalent fitting model and the kinetic parameters

obtained were ka1 = 1.5� 105 M � 1 s� 1, kd1 = 3.7�10� 5 s� 1 and KD1

app = 2.5� 10� 10 M (Rmax =

2400, v2 = 3.7). The 9C2 binding curve fitted

correctly with this model but gave a valid apparent

Fig. 3. Titration of AahI by a two-site immunometric assay. (A) and (B) standard curves were obtained for calibrated AahI toxin diluted in PBS

supplemented with 0.1% Tween and 0.1% casein (closed triangles) or in rabbit plasma (open triangles) or in human plasma (open circles) and

(C) kinetic profile of AahI in the plasma of envenomed rabbits. Venom (125 Ag/kg) was injected s.c. into rabbits, blood was collected at various

times after injection, and AahI concentration was quantified by ELISA. Experimental values were the mean of two independent experiments.


affinity constant only for the first site. The first

dissociation rate, corrected from a probable avidity

effect, still showed a slow dissociation rate.

3.4. Development of a two-site immunometric assay

Previous work using ELISA showed that only

mAb 9C2 and 2G3 could simultaneously bind to AahI

suggesting that these mAbs recognize two distinct

epitopes (Clot-Faybesse et al., 1999). To develop a

sensitive two-site immunoassay, protein A-purified

IgGs 9C2 and 2G3 were conjugated to biotin. Both

biotinylated mAbs conserved a high level of reactivity

with AahI, as shown by RIA (Fig. 2). These mAbs

were then used to set-up an ELISA for measuring

AahI concentration in venoms or fluids. The results

obtained in ELISA with the mAb 2G3/biotin-labeled

mAb 9C2 pair were not better than those obtained

with the mAb 9C2/biotin-labeled mAb 2G3 pair (data

not shown). We chose to use mAb 2G3 as capture

antibody because it was strictly specific for AahI.

Each assay parameter was changed until optimal

conditions were achieved (described in Materials

and methods). The titration curve of a standard AahI

solution is shown in Fig. 3A. The curve was clearly

linear between 0.04 and 10 ng/ml. The detection limit

of the assay was 75 pg/ml. The intra-assay coefficient

of variation was less than 1.3% and the inter-assay

coefficient of variation, as determined with five differ-

ent standard curves, was below 15%. Standard curves

were also constructed with the toxin diluted in non-

immune rabbit serum and in normal human plasma

(Fig. 3B). We then used this ELISA to measure the

AahI content of water-extracted venoms, collected

from various sites in Tunisia. Dilutions of up to 1 in

106 must be used due to the high sensitivity of the test.

According to our results, AahI accounted for 0.3–3%

of total venom proteins, depending on the origin of

the venom. Intra- and inter-assay coefficients of

variation were less than 5% and 10%, respectively.

Plasma AahI toxin concentrations were also measured

in experimentally envenomed rabbits. Fig. 3C shows

the changes in AahI concentration over time in the

plasma of a rabbit injected s.c. with sublethal doses of

Aah venom. AahI was detected soon after venom

injection and its concentration in plasma peaked

between 15 and 60 min decreasing progressively

thereafter to the detection limit of the test.

4. Discussion

4.1. Antibody response of mice to toxins and

production of mAbs

It has been noted that there is considerable varia-

tion in the immune responses of various inbred strains

of mice (Goding, 1980). As concern scorpion toxins,

mAbs against AahII have been obtained from BALB/c

mice (H-2d) immunized with native toxin (low-dose

protocols) (Bahraoui et al., 1988) or the nontoxic

analog (Abu)8 AahII (Devaux et al., 1997). In con-

trast, no significant antibody response was observed

in the serum of BALB/c mice injected with AahI. A

panel of 11 mouse strains of various MHC haplotypes

was initially tested and it was found that all strains

responded poorly (C57BL/6 mice occasionally gave a

positive signal in ELISA, but the spleens of these

mice were necrotic and the few growing hybrids, after

fusion, produced nonspecific IgM). It is therefore

clear that it may be difficult to raise mAbs against

pure toxins because of MHC restriction. This work

describes the trials performed with the aim of obtain-

ing high-affinity mAbs (IgG isotype) with various

specificities. The high toxicity displayed by scorpion

toxins made it impossible to inject large quantities of

immunogen; the useful doses were found to be in the

range of 4–12 Ag/mouse. We used standard protocols

(three to five injections) in C57BL/6 (H-2b) and

BALB/c (H-2d) mice, and tested various detoxified

immunogens. Injections of the nontoxic analog of

AahI ((Abu)8 AahI) or the glutaraldehyde-treated

AahI (injection doses of 10–50 Ag/mouse) induced

no detectable antibody response to AahI in either

strain of mice. AahI was then coupled to a carrier

protein by means of glutaraldehyde or carbodiimide.

This resulted in strong antibody responses, regardless

of the carrier protein or mouse strain used. The best

responses were obtained with BALB/c mice injected

with BSA-coupled AahI or with (Abu)8 AahII-gluta-

raldehyde-coupled AahI. Spleen cells from two of

these mice were used for hybridoma selection, and a

panel of high-affinity mAbs was produced (Clot-

Faybesse et al., 1999). These results suggest that

several points must be considered when trying to

produce mAbs directed against small molecules pre-

senting toxic activity in mammals. It is now well

known that a protein should possess certain molecular


features enabling it to trigger the various events

required to induce a humoral response and be a good

immunogen. This implies the presence, at the antigen

surface, of epitopes recognized by specific B-lympho-

cytes (B epitopes). The enzymatic processing of the

antigen by the cell machinery of the host is also

required to generate antigenic peptides (T epitopes),

which must bind to MHC molecules to be recognized

by T cell receptors (Lanzavecchia, 1985). This general

scheme probably applies to disulfide bridge-cross-

linked animal toxins. The immune response to apa-

min, an 18-amino acid molecule from bee venom, or

to snake neurotoxins (61 amino acids) have been

shown to be T cell dependent and to require the

modification of disulfide bridges before processing

by antigen-presenting cells (Regnier-Vigouroux et al.,

1988; Leonetti et al., 1990; Maillere et al., 1995).

Both antibody and T cell proliferation responses to

these toxins are under genetic control (Defendini et

al., 1988; Maillere et al., 1993). Although less well

documented, the Aah scorpion family of toxins acting

on sodium channels probably undergoes the same

kind of process. Aah scorpion toxins (about 7 kDa)

are large enough to induce an immune response. In

rabbits, high-quality polyclonal antibodies (Delori et

al., 1981) have been obtained following immunization

with pure toxins, but the titers of anti-AahII pAb

(1:150,000) were higher than those of anti-AahI

(1:30,000) and anti-AahIII (1:10,000) pAbs (C.

Devaux, unpublished data). In mice, AahII and its

synthetic analog (Abu)8 AahII induced an adequate

immune response implying that both the toxin and its

analog contain B and T epitopes. The lack of response

to AahI observed in several mouse strains cannot be

attributed to the size of the immunogen (similar to that

of AahII) or to the residual toxicity of the immunogen

(toxic effect by the s.c. route: 180 and 360 ng/20 g

mouse for AahI and AahII, respectively). Likely, the

antibody response to AahI depends on appropriate

presentation by H-2b or H-2u MHC molecules of as

yet undefined T cell epitopes. An adequate anti-AahI

response was finally obtained by coupling the toxin to

a carrier protein, a procedure often used for synthetic

low-molecular weight peptides (Briand et al., 1985;

Muller, 1988). We identified three useful carrier

proteins: the widely used proteins BSA (67 kDa)

and KLH (about 2000 kDa), and a much smaller

peptide, the synthetic (Abu)8 AahII (7.2 kDa). This

result confirms that the entire primary sequence of

AahII, despite the lack of disulfide bridges, contains

all the information necessary to induce an immune

response (Devaux et al., 1997; Zenouaki et al., 1997).

Our results support the view that this is not the case

for AahI. The amino acid sequences of AahI and

AahII differ by 56% (Rochat et al., 1979). Thus, one

possible cause of the lack of response to AahI in mice

is sequence differences, which may alter the process-

ing or presentation of the antigen. This may in turn

lead to the loss or modification of T cell epitopes

rendering them unable to bind MHC molecules

(Moudgil et al., 1998).

4.2. Characterization and use of mAbs anti-toxins

The anti-AahI mAbs we obtained had several

characteristics rendering them of potential value both

for anti-venom therapy and research use. All the

selected mAbs displayed high affinity for the cognate

toxin; the KD values obtained in this work by BIA-

CORE being consistent with those deduced by RIA

(Clot-Faybesse et al., 1999; Devaux et al., 2001). One

of them, mAb 9C2, neutralized AahI and AahIII, two

toxins belonging to the same immunological group,

and reduced the toxicity of the whole venom. This

reduction was enhanced when the anti-AahI mAb 9C2

was coinjected into mice with the anti-AahII mAb

4C1 (Bahraoui et al., 1988; Clot-Faybesse et al.,

1999). These neutralizing properties were probably

due to an overlap between the epitope of mAb 9C2

and the pharmacological site on AahI, and are prob-

ably strengthened by the high affinity of mAb 9C2 for

AahI, as determined by BIACORE technology

(KD = 0.11�10� 10 M), due to a slow dissociation

rate.

An adequate treatment of envenoming requires

accurate evaluation of toxin concentration in plasma

so as to follow toxicokinetic parameters in victims. To

date, only polyclonal antibody-based RIA has been

shown to be sensitive enough to determine individual

neurotoxins in scorpion venoms or in sera (reviewed

by Devaux and El Ayeb, 2000). However, RIA

involves the handling of radiolabeled molecules and

is not suitable for use in the emergency conditions of a

clinical diagnosis or therapy in envenomed patients.

We used two anti-AahI mAbs to develop a specific

two-site immunoassay for AahI by ELISA. The bio-


tin/peroxidase-labeled streptavidin amplification sys-

tem (Diamandis and Christopoulos, 1991) was used

and yielded an assay of a sensitivity similar to that

reached in RIA (0.1 ng/ml). No cross-reactivity was

observed with toxins from the same immunological

group (AahIII) or of another group (AahII) up to the

highest toxin concentration tested (40 ng/ml). We used

this assay, first, to determine AahI concentration in the

highly diluted venoms of scorpions from five different

locations, and second, to follow the pharmacokinetics

of AahI toxin in experimentally envenomed rabbits.

The amount of AahI in venoms (obtained by manual

sampling) was variable, as previously suggested (El

Ayeb and Rochat, 1985), ranging from 0.3% to 3%.

We also found that AahI was detected very soon after

venom injection into rabbits, with the maximum

concentration reached after about 1 h, followed by a

progressive decrease until 8 h. Although an ELISA for

the determination of scorpion venom in sera has been

described (Chavez-Olortegui et al., 1994; Krifi et al.,

1998), this study is the first to report the specific

determination of an individual toxin in envenomed

animal fluids. Work is underway to develop a sim-

ilarly sensitive ELISA for the determination of AahII,

the other toxin responsible for deaths due to enve-

noming. The ELISA described here is sufficiently

sensitive for determining the fate of AahI directly in

sera from envenomed humans. It could be used for

retrospective studies, thereby providing a useful tool

for laboratory investigation of envenoming cases.

In conclusion, this work shows that toxin AahI

must be attached to a helper molecule if it is to

induce an immune response in mice. Using this

strategy, we produced a panel of high-affinity mAbs

specific for the toxin. An ELISA using two of these

mAbs was designed to specifically determine AahI

over a range of concentrations similar to that covered

by RIA and appropriate for measurements in bio-

logical fluids.

Acknowledgements

We would like to thank Dr. Jean-Marc Sabatier for

providing anatoxins, Dr. Mohamed N. Krifi for rabbit

management in pharmacokinetic studies, and Dr.

Maria Leria Defendini for her previous work on

mouse immunization. We also thank Emilie Flament

and Marianick Juin for technical assistance and

Maryse Alvytre for animal care.

References

Amitai, Y., 1998. Clinical manifestations and management of scor-

pion envenomation. Public Health Rev. 26, 257–263.

Bahraoui, E., Pichon, J., Muller, J.M., Darbon, H., El Ayeb, M.,

Granier, C., Marvaldi, J., Rochat, H., 1988. Monoclonal anti-

bodies to scorpion toxins. Characterization and molecular mech-

anisms of neutralization. J. Immunol. 141, 214–220.

Briand, J.P., Muller, S., Van Regenmortel, M.H., 1985. Synthetic

peptides as antigens: pitfalls of conjugation methods. J. Immu-

nol. Methods 78, 59–69.

Chavez-Olortegui, C., Fonseca, S.C., Campolina, D., Amaral, C.F.,

Diniz, C.R., 1994. ELISA for the detection of toxic antigens in

experimental and clinical envenoming by Tityus serrulatus scor-

pion venom. Toxicon 32, 1649–1656.

Clot-Faybesse, O., Juin, M., Rochat, H., Devaux, C., 1999. Mono-

clonal antibodies against the Androctonus australis hector scor-

pion neurotoxin I: characterisation and use for venom neutrali-

sation. FEBS Lett. 458, 313–318.

Defendini, M.L., El Ayeb, M., Regnier-Vigouroux, A., Granier, C.,

Pierres, M., 1988. H-2A-linked control of T-cell and antibody

responses to apamin. Immunogenetics 28, 139–141.

Dehesa-Davila, M., Possani, L.D., 1994. Scorpionism and serother-

apy in Mexico. Toxicon 32, 1015–1018.

Delori, P., Van Rietschoten, J., Rochat, H., 1981. Scorpion venoms

and neurotoxins: an immunological study. Toxicon 19, 393–407.

Devaux, C., El Ayeb, M., 2000. Immunological properties of scor-

pion toxins. In: Rochat, H., Martin-Eauclaire, M.-F. (Eds.),

Animal Toxins, Facts and Protocols. Birkhauser Verlag, Basel,

pp. 183–195.

Devaux, C., Clot-Faybesse, O., Juin, M., Mabrouk, K., Sabatier

J.M., Rochat, H., 1997. Monoclonal antibodies neutralizing

the toxin II from Androctonus australis hector scorpion venom:

usefulness of a synthetic, nontoxic analog. FEBS Lett. 412,

456–460.

Devaux, C., Moreau, E., Goyffon, M., Rochat, H., Billiald, P., 2001.

Construction and functional evaluation of a single-chain anti-

body fragment that neutralizes toxin AahI from the venom of the

scorpion Androctonus australis hector. Eur. J. Biochem. 268,

694–702.

Diamandis, E.P., Christopoulos, T.K., 1991. The biotin– (strept)avi-

din system: principles and applications in biotechnology. Clin.

Chem. 37, 625–636.

El Ayeb, M., Rochat, H., 1985. Polymorphism and quantitative

variations of toxins in the venom of the scorpion Androctonus

australis hector. Toxicon 23, 755–760.

El Ayeb, M., Delori, P., Rochat, H., 1983. Immunochemistry of

scorpion alpha-toxins: antigenic homologies checked with ra-

dioimmunoassays (RIA). Toxicon 21, 709–716.

Goding, J.W., 1980. Antibody production by hybridomas. J. Immu-

nol. Methods 39, 285–308.

Goyffon, M., Vachon, M., Broglio, N., 1982. Epidemiological and


clinical characteristics of the scorpion envenomation in Tunisia.

Toxicon 20, 337–344.

Heard, K., O’Malley, G.F., Dart, R.C., 1999. Antivenom therapy in

the Americas. Drugs 58, 5–15.

Ho, M., Warrell, M.J., Warrell, D.A., Bidwell, D., Voller, A., 1986.

A critical reappraisal of the use of enzyme-linked immunosorb-

ent assays in the study of snake bite. Toxicon 24, 211–221.

Kharrat, R., Zenouaki, I., Ben Lasfar, Z., Miled, K., El Ayeb, M.,

1997. Molecular characterization, antigenicity and immunoge-

nicity of anatoxic polymeric forms conferring protection against

scorpion venoms. Toxicon 35, 915–930.

Krifi, M.N., Kharrat, H., Zghal, K., Abdouli, M., Abroug, F., Bou-

choucha, S., Dellagi, K., El Ayeb, M., 1998. Development of an

ELISA for the detection of scorpion venoms in sera of humans

envenomed by Androctonus australis garzonii (Aag) and Buthus

occitanus tunetanus (Bot): correlation with clinical severity of

envenoming in Tunisia. Toxicon 36, 887–900.

Krifi, M.N., Amri, F., Kharrat, H., El Ayeb, M., 1999. Valuation of

antivenom therapy in children severely envenomed by Androc-

tonus australis garzonii (Aag) and Buthus occitanus tunetanus

(Bot) scorpions. Toxicon 37, 1627–1634.

Lanzavecchia, A., 1985. Antigen-specific interaction between T and

B cells. Nature 314, 537–539.

Leonetti, M., Pillet, L., Maillere, B., Lamthanh, H., Frachon, P.,

Couderc, J., Menez, A., 1990. Immunization with a peptide

having both T cell and conformationally restricted B cell epito-

pes elicits neutralizing antisera against a snake neurotoxin. J.

Immunol. 145, 4214–4221.

Maillere, B., Cotton, J., Mourier, G., Leonetti, M., Leroy, S., Me-

nez, A., 1993. Role of thiols in the presentation of a snake toxin

to murine T cells. J. Immunol. 150, 5270–5280.

Maillere, B., Mourier, G., Herve, M., Cotton, J., Leroy, S., Menez,

A., 1995. Immunogenicity of a disulphide-containing neurotox-

in: presentation to T-cells requires a reduction step. Toxicon 33,

475–482.

Miranda, F., Kopeyan, C., Rochat, H., Rochat, C., Lissitzky, S.,

1970. Purification of animal neurotoxins. Isolation and charac-

terization of eleven neurotoxins from the venoms of the scor-

pions Androctonus australis hector, Buthus occitanus tunetanus

and Leiurus quinquestriatus quinquestriatus. Eur. J. Biochem.

16, 514–523.

Moudgil, K.D., Sercarz, E.E., Grewal, I.S., 1998. Modulation of the

immunogenicity of antigenic determinants by their flanking res-

idues. Immunol. Today 19, 217–220.

Muller, S., 1988. Peptide-carrier conjugaison. In: Burdon, R.H., van

Knippenberg, P.H. (Eds.), Synthetic Peptides as Antigens in

Laboratory Techniques in Biochemistry and Molecular Biology,

vol. 19. Elsevier, Amsterdam, pp. 95–127.

Possani, L.D., Fernandez de Castro, J., Julia, J.Z., 1981. Detoxifi-

cation with glutaraldehyde of purified scorpion (Centruroides

noxius hoffmann) venom. Toxicon 19, 323–329.

Regnier-Vigouroux, A., El Ayeb, M., Defendini, M.L., Granier, C.,

Pierres, M., 1988. Processing by accessory cells for presentation

to murine T cells of apamin, a disulfide-bonded 18 amino acid

peptide. J. Immunol. 140, 1069–1075.

Rochat, H., Tessier, M., Miranda, F., Lissitzky, S., 1977. Radio-

iodination of scorpion and snake toxins. Anal. Biochem. 82,

532–548.

Rochat, H., Bernard, P., Couraud, F., 1979. Scorpion toxins: chem-

istry and mode of action. Adv. Cytopharmacol. 3, 325–334.

Zenouaki, I., Kharrat, R., Sabatier, J.M., Devaux, C., Karoui, H.,

Van Rietschoten, J., El Ayeb, M., Rochat, H., 1997. In vivo

protection against Androctonus australis hector scorpion toxin

and venom by immunization with a synthetic analog of toxin II.

Vaccine 15, 187–194.


a strategy for inducing an immune response against androctonus australis scorpion venom toxin i in...

Documents