WO2009150539A2 - Variable domains of camelid heavy-chain antibodies directed against androctonus autralis hector toxins - Google Patents

Abstract

The present invention relates to variable domains of camelid heavy- chain antibodies directed against Androctonus australis hector (Aah) toxins and uses thereof for preparing therapeutic or diagnostic agents.

Description

VARIABLE DOMAINS OF CAMELID HEAVY-CHAIN ANTIBODIES DIRECTED AGAINST ANDROCTONUS AUSTRALIS HECTOR TOXINS

The present invention relates to antibodies directed against Androctonus australis hector (Aah) toxins, and their use for treating envenomed patients. Particularly, these antibodies (or binders) are derived from camelid Heavy- chain antibodies (HCAb).

Humans in tropical and subtropical regions are heavily exposed to hazardous scorpion stings. For example, 200,000 epidemiological cases are reported annually in Mexico [1] and around 30,000 in Tunisia (27,000 for 2006, 'Direction de soin et de sante de base, DSSB, Ministere de Ia sante') of which 1,000 have systemic manifestations requiring admission to hospital [2-3]. For the latter country, the scorpion Androctonus australis hector (Aah) is responsible for most of the registered fatal stings, thereby constituting a serious health problem [3].

The molecules responsible for the noxious effects of scorpion venoms are polypeptides of about 65 amino acid residues. These small proteins dramatically affect excitable cells by specific interaction with voltage gated sensitive sodium channels [4-5]. Toxins comprising less than 2% of venom dry weight were purified from animals collected in Algeria and sequenced: the toxins Aahl [49], AahF [49], Aahl" [49], AahII [50] and AahIII [51] are active on mammals, and the toxin AahIT [52] is active on insects (see for review: Guidebook to Protein Toxins and Their Use in Cell Biology, Edited by Rino Rappuoli and Cesare Montecucco, 278 p., JuI. 1997). The Aahl' differs by one amino acid from Aahl to which it is otherwise antigenically identical and it is structurally related to AahII with which it does not share antibody cross-reactivity [53, 12]. The venom of animals collected in Tunisia (area of Tozeur) differs from the one of the Algeria animals by the absence of Aahl" and the presence of a second toxin active on insects: AahIT2 [54].

The most abundant toxic compound of the Aah venom [6] referred to as AahII has been purified [7-8]. The AahII is also the most poisonous toxin of this North African scorpion [6] with an LDso < 3 ng; its structural and antigenic properties are well established [9-13].

The current immunotherapeutic treatment of scorpion envenoming consists in administering purified polyclonal F(ab')2 fractions prepared from equine hyperimmune sera [3, 14]. Their manufacturing is tedious and sometimes F(ab')2 fragments may still exert adverse effects upon injection in the patient, e.g. anaphylactic shock [15-17]. In short, the antivenom in scorpion sting treatment is - at least - controversial. In Brazil [18] and Mexico [1] its effectiveness in reducing the pediatric scorpion sting mortality is beyond any doubt, whereas in India [19], Israel [20] and Tunisia [3, 17] the beneficial effects of routine administration of scorpion antivenom to stung patients are uncertain. In addition, in severe cases of scorpion envenoming where the toxins diffuse rapidly from the blood, the administered F(ab')2 based antivenom might compete inefficiently with Na* channels for capturing the toxins. The reason for this low efficacy in competing is ascribed - at least in part - to the large difference in molecular weight (and diffusion), between the F(ab')2 and the scorpion toxin polypeptides [15].

It appears from the foregoing that there is room for improvement on the currently used anti-scorpion venom therapy. The introduction of smaller, recombinant antigen-binding fragments derived from monoclonal antibodies is considered to be such an improvement on the current therapy. Indeed, one Aahll-specific murine monoclonal antibody (4Cl) with neutralizing capacity has been isolated [21] and was subsequently employed as starting material to develop a single chain variable fragment (scFv) exhibiting a high affinity towards AahII [22]. A second scFv, against the Aahl toxin was identified [23]; further engineering led to the design of a bispecific scFv construct against the Aahl and AahII toxins that protected mice against the Aah venom [24]. However, this bispecific antibody construct (MW of about 60 kDa) would be of limited use to face the rapid kinetic diffusion of the toxin (MW of about 7 kDa). Moreover, their bacterial production levels are relatively low, the neutralizing capacity remains moderate and its use in human therapeutic might still generate an undesirable human anti-mouse antibody (HAMA) response. To avoid a possible HAMA response against the antivenom, human scFv antibody fragments were generated against the Cn2 toxin, the major toxic component of the new world scorpion Centruroides noxius [25]. The subsequent development of recombinant Fabs from these scFv were shown to possess improved stability and toxin neutralization properties [26]. Within the framework of research that has led to the present invention, the Inventors have obtained camelid homodimeric Heavy-chain antibodies (HCAbs, [27]) directed against Aah toxins, prepared VHH derived therefrom and analysed their properties both in vitro and in vivo. The VHH domain (referred to as VHH, VHH binder or Nanobody), derived from HCAbs, represents the smallest, intact, natural antigen-binding fragment with a MW of only about 15 kDa. The small size (in the nm range), good stability, high level of expression in microbial systems, high solubility, good specificity, high affinity of the Nanobodies (Nbs, [41]) for their cognate antigen and their close sequence identity to human VH of family III makes them potentially valuable for therapeutic applications [27-30].

In a previous study, the Inventors have demonstrated the possibility to elicit HCAb immune response in a dromedary that could neutralize the AahG50 venom fraction containing Aahl' and AahII [31]. The Inventors have now generated a VHH library from AahG50-immunized dromedary and Nanobodies (also named binders) directed to Aahl' were selected by phage display. One of these Nanobodies (named NbAahI'22) was used to reconstruct a functional HCAb and to generate a tandem linked bivalent Nanobody. The Inventors have observed that the monomeric and dimeric bivalent Nanobody exhibited an Aahl' toxin neutralizing capacity exceeding that of previously tested scFv-based constructs [23, 24, 42].

The Inventors have also generated an immune VHH library from the lymphocytes of a dromedary that was immunized with purified AahII, and retrieved after phage display multiple Nanobodies that recognize the AahII with high affinity. From all the recombinant antibodies, some of the Aahll-specific Nanabodies showed a very good scorpion toxin-neutralizing activity in mice.

Accordingly, the present invention provides an isolated variable domain of a camelid heavy-chain antibody (VHH domain) directed against an Androctonus australis hector (Aah) toxin or the isolated CDR3 region therefrom.

A VHH domain refers usually to a variable domain of a camelid (dromedary, camel, llama, alpaca,...) heavy-chain antibody. As used herein, a VHH domain can also be named "binder" or Nanobody (Nb). The term "isolated" refers to a VHH domain which has been separated from a camelid heavy-chain antibody from which it derives.

According to the present invention, a VHH domain comprises a recombinant or synthetic VHH domain. The term "recombinant" refers to the use of genetic engineering methods (cloning, amplification) to produce said VHH domain.

The term "synthetic" refers to production by in vitro chemical or enzymatic synthesis.

Preferably, the VHH domain of the invention is isolated from a dromedary heavy-chain antibody.

Preferably, the VHH domain of the invention consists of 1 10 to 150 amino acid residues.

An Androctonus australis hector (Aah) toxin refers to a toxin protein from Androctonus australis hector scorpion. Preferably said Aah toxin is selected from the protein toxins Aahl [49], AaW' [49], Aahl" [49], AahII [50], AahIII [51], AahIT [52] and AahIT2 [54]. More preferably, said Aah toxin is Aahl' or AahII.

In a preferred embodiment of the invention, the VHH domain is directed against Aahl' and is selected from the group consisting of AahI'B2 (SEQ ID NO: 1), Aahl'Bl (SEQ ID NO: 2), AahI'A17 (SEQ ID NO: 3), AaW'A3 (SEQ ID NO: 4), AahI'F12 (SEQ ID NO: 5), AaW'D19 (SEQ ID NO: 6), AahI'DIO (SEQ ID NO: 7), Aahl'CH (SEQ ID NO: 8), AahI'C13 (SEQ ID NO: 9), AaW'B23 (SEQ ID NO: 10), Aahl'Dl l (SEQ ID NO: 1 1), AahI'A2 (SEQ ID NO: 12), AahI'A19 (SEQ ID NO: 13) and NbAahI'22 (SEQ ID NO: 1 1 1) in reference to figures 4 and 5.

The CDR3 regions from the VHH domains represented as SEQ ID NO: 1 to 13 are respectively represented herein as SEQ ID NO: 14 to 26 in reference to Table 4.

In a more preferred embodiment of the invention, the VHH domain directed against AaW is AahI'F12 (SEQ ID NO: 5) and the CDR3 region therefrom is SEQ ID NO: 18 or NbAaW'22 (SEQ ID NO: 1 1 1) and the CDR3 region therefrom. In another preferred embodiment of the invention, the VHH domain is directed against AahII and is selected from the group consisting of NbAahII47 (SEQ ID NO: 27), NbAahIIlbis (SEQ ID NO: 28), NbAahII23 (SEQ ID NO: 29), NbAahII30 (SEQ ID NO: 30), NbAahII33 (SEQ ID NO: 31), NbAahII15 (SEQ ID NO: 32), NbAahlllό (SEQ ID NO: 33), NbAahII36 (SEQ ID NO: 34), NbAahID (SEQ ID NO: 35), NbAahllό (SEQ ID NO: 36), NbAahII17 (SEQ ID NO: 37), NbAaWI 1 1 (SEQ ID NO: 38), NbAahII32 (SEQ ID NO: 39), NbAahIIlO (SEQ ID NO: 40), NbAahII2bis (SEQ ID NO: 41), NbAahII9 (SEQ ID NO: 42), NbAahII08 (SEQ ID NO: 43), NbAahII18 (SEQ ID NO: 44), NbAahII45 (SEQ ID NO: 45), NbAahII05 (SEQ ID NO: 46), NbAahII48 (SEQ ID NO: 47), NbAahII39 (SEQ ID NO: 48), NbAahII38 (SEQ ID NO: 49), NbAahII29b (SEQ ID NO: 50), NbAahII37bis (SEQ ID NO: 51), NbAahII41 (SEQ ID NO: 52), NbAahII43 (SEQ ID NO: 53), NbAahII20 (SEQ ID NO: 54), NbAahII21 (SEQ ID NO: 55), NbAahII44 (SEQ ID NO: 56), NbAahII4bis (SEQ ID NO: 57), NbAahII28bis (SEQ ID NO: 58), NbAahII12 (SEQ ID NO: 59), NbAahII13 (SEQ ID NO: 60), NbAahII14bis (SEQ ID NO: 61), NbAahII35 (SEQ ID NO: 62) and NbAahII19 (SEQ ID NO: 63) in reference to figure 1. The CDR3 regions from these VHH domains are respectively represented herein as SEQ ID NO: 64-100 in reference to figure 1.

In a more preferred embodiment of the invention, the VHH domain directed against AahII is NbAahIIlO (SEQ ID NO: 40) and the CDR3 region therefrom is represented as SEQ ID NO: 77. A VHH domain of the invention is obtainable by the method comprising the steps of:

(a) immunizing a camelid, preferably a dromedary, with a Aah toxin, such as Aahl' or AahII, preferably purified or a mixture of Aah toxins, preferably purified such as the AahG50 fraction, (b) isolating peripheral lymphocytes of the immunized camelid, obtaining the total RNA and synthesizing the corresponding cDNAs (methods are known in the art);

(c) constructing a library of cDNA fragments encoding VHH domains, (d) transcribing the VHH domain-encoding cDNAs obtained in step

(c) to mRNA using PCR, converting the mRNA to phage display format, and selecting the VHH domain by phage display, and (e) expressing the VHH domain in a vector (for instance, a suitable vector is the phagemid vector pHENό [33]) and, optionally purifying the expressed VHH domain.

The present invention also provides an isolated polypeptide comprising at lest one VHH domain or at least one CDR3 region as defined above.

Such a polypeptide can be a fusion protein between a human Fc antibody fragment and at least one VHH domain or at least one CDR3 region as defined above.

When the polypeptide of the present invention comprises at least two VHH domains or CDR3 regions as defined above, then said VHH domains or CDR3 regions can be identical (homomultimer) or different (heteromultimer) and can be separated from one another by a spacer, preferably an amino acid spacer.

When the polypeptide of the present invention comprises at least two different VHH domains or CDR3 regions as defined above, then said VHH domains or CDR3 regions can be directed against the same or a different Aah toxin.

The present invention also provides an isolated polynucleotide encoding a VHH domain, CDR3 region or polypeptide of the present invention. The polynucleotides of the invention may be obtained by the well-known methods of recombinant DNA technology and/or of chemical DNA synthesis. By way of example, a polynucleotide encoding a peptide comprising

NbAahI'22 is represented as SEQ ID NO: 1 10.

The present invention also provides recombinant expression cassettes comprising a polynucleotide of the invention under the control of a transcriptional promoter allowing the regulation of the transcription of said polynucleotide in a host cell. Said polynucleotide can also be linked to appropriate control sequences allowing the regulation of its translation in a host cell.

The present invention also provides recombinant vectors comprising a polynucleotide or an expression cassette of the invention.

The present invention also provides a host cell containing a recombinant expression cassette or a recombinant vector of the invention. The host cell is either a prokaryotic or eukaryotic host cell. The present invention also provides a therapeutic or diagnostic agent comprising a VHH domain, CDR3 region or polypeptide of the present invention.

When the VHH domain, CDR3 region or polypeptide of the present invention is administered to a human subject, then they can be humanized to reduce immunogenicity in human. Methods for producing humanized antibodies or fragments thereof are known in the art.

In an embodiment of said diagnostic agent, the VHH domain, CDR3 region or polypeptide of the invention can be linked, directly or indirectly, covalently or non-covalently to a diagnostic compound. The diagnostic compound can be directly and covalently linked to the VHH domain, CDR3 region or polypeptide of the present invention either to one of the terminal ends (N or C terminus) of said VHH domain, CDR3 region or polypeptide, or to the side chain of one of the amino acids of said VHH domain, CDR3 region or polypeptide. The diagnostic compound can also be indirectly and covalently linked to said VHH domain, CDR3 region or polypeptide by a connecting arm (i.e., a cross-linking reagent) either to one of the terminal ends of said VHH domain or polypeptide or to a side chain of one of the amino acids of said VHH domain or polypeptide. Linking methods of a compound of interest to a peptide are well known in the art. In a preferred embodiment of said diagnostic agent, said diagnostic compound is selected from the group consisting of:

- enzymes such as horseradish peroxidase, alkaline phosphatase, glucose-6-phosphatase or beta-galactosidase;

- fluorophores such as green fluorescent protein (GFP), blue fluorescent dyes excited at wavelengths in the ultraviolet (UV) part of the spectrum

(e.g. AMCA (7-amino-4-methylcoumarin-3-acetic acid); Alexa Fluor 350), green fluorescent dyes excited by blue light (e.g. FITC, Cy2, Alexa Fluor 488), red fluorescent dyes excited by green light (e.g. rhodamines, Texas Red, Cy3, Alexa Fluor dyes 546, 564 and 594), or dyes excited with far-red light (e.g. Cy5) to be visualized with electronic detectors (CCD cameras, photomultipliers);

- heavy metal chelates such as europium, lanthanum or yttrium; - radioisotopes such as [¹⁸F]fluorodeoxy glucose, ^{1 1}C-, ¹²⁵I-, ¹³¹I-, ³H-, ¹⁴C-, ³⁵S, or ⁹⁹Tc- labelled compounds.

The present invention also provides a kit for diagnosing or monitoring, in a subject, an envenoming and an intoxication by Androctonus australis hector, comprising at least a VHH domain, a CDR3 region, a polypeptide, a diagnostic agent or a polynucleotide of the present invention.

The VHH domain, the CDR3 region, the polypeptide, the therapeutic or diagnostic agent, or the polynucleotide of the present invention can be administered to a subject (a mammal, and preferably a human) by injection, such as intravenous, intraperitoneal, intramuscular or subcutaneous injection.

The present invention also provides a method for diagnosing a subject envenomed by Androctonus australis hector, comprising the steps of: a) contacting in vitro or ex vivo an appropriate biological sample from said subject with a VHH domain, CDR3 region, polypeptide or a diagnostic agent of the present invention, b) determining the presence or the absence of an Aah toxin in said biological sample, the presence of said Aah toxin indicating that said subject is envenomed by Androctonus australis hector.

The present invention also provides a pharmaceutical composition comprising a therapeutic agent as defined above and a pharmaceutically acceptable carrier.

As used herein, "pharmaceutically acceptable carrier" is intended to include any and all solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic and absorption delaying agents, and the like, compatible with pharma- ceutical administration. Suitable carriers are described in the most recent edition of

Remington's Pharmaceutical Sciences, a standard reference text in the field. Preferred examples of such carriers or diluents include, but are not limited to, water, saline,

Ringer's solutions, dextrose solution, and 5% human serum albumin. Liposomes, cationic lipids and non-aqueous vehicles such as fixed oils may also be used. The use of such media and agents for pharmaceutically active substances is well known in the art. Except insofar as any conventional media or agent is incompatible with a therapeutic agent as defined hereabove, use thereof in the composition of the present invention is contemplated.

The present invention also provides a VHH domain, a CDR3 region, a polypeptide, a therapeutic agent, a pharmaceutical composition or a polynucleotide of the present invention for use in the treatment or prevention of envenoming and intoxication by Androctonus australis hector.

As used herein, the term "treatment" includes the administration of the VHH domain, CDR3 region, polypeptide, polynucleotide, therapeutic agent or a pharmaceutical composition as defined above to a patient who is envenomed and intoxicated by Androctonus australis hector, with the purpose to cure, heal, alleviate, relieve, alter, remedy, ameliorate, improve or affect the envenoming or intoxication, or the symptoms of the envenoming or intoxication.

As used herein, the term "prevention" includes the inhibition or the averting of symptoms associated with envenoming or intoxication by Androctonus australis hector.

In addition to the preceding features, the invention further comprises other features which will emerge from the following description, which refers to examples illustrating the present invention, as well as to the appended figures.

Figure 1 represents the alignment of the deduced amino acid sequences of the AahII specific VHH domains (Nanobodies, Nbs) numbered according to the ImMunoGeneTics numbering (IMGT) [36]. Framework domains and hypervariable domains (CDR) are indicated at the top of the figure. Letters in front of the sequence refers to the various clusters of the Nbs according to their CDR3 homology. The Nbs evaluated for .their in vivo toxin neutralization are highlighted in grey. Dots indicate an identical amino acid at that position relative to the top sequence within that cluster, dashes are introduced to align the sequences.

Figure 2 represents the surface plasmon resonance sensograms whereby AahII is fixed on the sensor layer and Nbs directed against this toxin are in the mobile phase. (A) Analysis whereby different concentrations of NbAahII lO (as indicated) are injected (at time = 290 s) at a flow-rate of 30 μl/min to the AahII toxin on the chip, the bound Nb is washed away with buffer (from 470 s onwards). (B) Epitope mapping whereby an excess of the monoclonal NbAahII 10 is first injected (at 215 s) to saturate its epitope, and from 330 s onwards a mixture of NbAahIIlO (SEQ ID NO:40) with NbAahII05 (SEQ ID NO:46) (as indicated) is applied. A net signal increase is observed for all Nbs (except NbAahIIlO) indicating that they are binding to an epitope that is non-overlapping with the NbAahIIlO epitope. Figure 3 is a schematic representation of the epitope complementation groups. After performing the epitope mapping of the Nbs directed against AahII with all possible mixtures and sequential additions (one example with NbAahIIlO as first Nb is shown in figure 2B) the complementation groups are inferred. The epitopes of the examples of the various clusters are denoted by circles, the Nb AahII clone number and the cluster to which this clone belongs are given. It seems that clones NbAahII20 and -12 of cluster I bind to an identical epitope, whereas clones NbAahII17 and 38 of clusters D and H respectively bind to the same epitope complementation group. The epitope of NbAahIIlO does not overlap with epitopes from other Nbs, however, the exact location of this epitope remains therefore tentative. The black, grey and white colours of the circles refer to strong, intermediate and weak neutralization of AahII toxin, respectively.

Figure 4 represents the alignment of different VHH domains (binders) directed against Aahl', according to the IMGT amino acid numbering.

Figure 5 represents a nucleotide sequence (SEQ ID NO: 1 10) and deduced amino acid sequence (SEQ ID NO: 1 1 1) comprising NbAahI'22 (amino acids 1-128), according to the IMGT amino acid numbering. Framework regions are in bold. The Complementary Determining Regions are underlined. The * corresponds to the STOP codon. The VHH-hallmark amino acid at positions 12, 42, 49, 50 and 52 are in italic. Figure 6 is a graphic showing the serum-stability of the monomeric

Nanobody NbAahI'22 (rhomb), the tandem linked bivalent Nb format (square) and the chimeric Nb-human Fc bivalent (triangle) incubated for various time periods (1 , 20, 25, 44, 49 h) in human serum at 37°C before testing the residual antigen binding capacity by ELISA. Figure 7 represents the temperature induced unfolding of monovalent (round) and bivalent (square) NbAahI'22 Nanobodies. Fraction of VHH unfolded, βi as function of temperature (T, ⁰C). EXAMPLE 1: EXPERIMENTAL PROCEDURES Animals

- Pubertal female dromedary (Camelus dromedaήus) weighting 230 kg was kept at the region of Kebili, Tunisia. - Swiss mice, 8 weeks old, weighting (20 ± 2) g were provided by the 'Service des Unites Animalieres' of the 'Institut Pasteur de Tunis'.

- Androctonus australis hector scorpions were collected in the Southern area of Tunisia.

Venom and toxins The Androctonus australis hector scorpions were electrically stimulated to release venom. The collected venom was subsequently purified on a Sephadex G50 as previously described [31]. The whole toxic fraction was named AahG50. Applying this AahG50 fraction on a Mono-S FPLC column (5/5 Mono-S, Pharmacia) and using a 35 min gradient from 0 to 60 % of Ammonium Acetate (of 0.05 M and 0.5 M), yielded a highly enriched AahII and Aahl/Aahl' fraction (eluting at 45% of the gradient) that is recognized by AahII- and Aahl/Aahl' specific rabbit polyclonal antibodies. This highly enriched, antigenically recognized fraction was further purified on a reversed phase HPLC C8 column (Beckman) in water with 0.1% Trifluoroacetic acid, using a 30 min gradient from 20 to 50 % of solution B (0.1% Trifluoroacetic acid in acetonitrile). The N-terminal sequence (Edman degradation) and MALDI-TOF mass spectrometry of purified AahII and Aahl' were controlled and shown to correspond to the expected sequence and mass.

Aahl and its homologue Aahl' differ only by a single amino acid substitution Val l7Ile. MALDI-TOF mass spectrometry was used to discriminate the Aahl and Aahl' proteins as described by Delabre et al. [55]

The toxicity was assessed by intracerebroventricular (i.c.v.) injection in Swiss mice (see below).

Escherichia coli strains and vectors

The E. coli strain TGl was used to host the VHH library. The phage display vector pHEN4 was employed to construct the VHH library as described by

Arbabi-Ghahroudi et al. [32]. The Nb genes selected by biopanning were recloned in the expression vector pHEN6 [33], This vector encodes a PeIB leader signal peptide to secrete the Nb in the periplasmic compartment of bacteria, and a C-terminal His6 for affinity purification and detection of the recombinant Nb. The E.coli strain WK6 was used for the Nb gene expression [33].

Dromedary immunization To obtain HCAb directed against Aahll, a female, healthy dromedary (Camelus dromedarius) of 3 years old was injected subcutaneously, five times (at days 0, 7, 14, 21 and 35) with an increasing amount of the Aahll enriched fraction and then with the purified Aahll toxin (80 to 100 μg at days 49, 57 and 64). To maintain the high titre of immunoglobulins (and antigen-specific B cells), the dromedary received three additional boosts of Aahll (100 μg at days 94, 101 and 108). The first injection was in Complete Freund's Adjuvant and booster injections were in Incomplete Freud's Adjuvant emulsions.

To obtain HCAb directed against Aahl', a dromedary was injected several times subcutaneously with increasing amount of AahG50 toxic fraction as described by Meddeb-Mouehli et al. [31]

Construction of an 'immune' VHH library Four days after the last boost, over 100 ml of anti-coagulated blood was collected from which plasma and peripheral blood lymphocytes were isolated by centrifugation on Lymphoprep (Nycomed). Total RNA from approximately 108 lymphocytes was released in 4 ml of a solution containing 23.6 g guanidine-SCN, 0.37 g citric acid, 0.25 g lauroyl sarcosine for 50 ml RNase-free water and adjusted to pH 7.0. After shearing by pushing through 19G and 23 G needles, 400 μl 2M Na-acetate (pH 4.0), 4 ml water saturated phenol and 2 ml chloroform/isoamylalcohol (24/1) were added to extract the RNA in the aqueous phase. After ethanol precipitation and dissolving the RNA in 300 μl RNase-free water, 50 μg is used to prepare cDNA with 2.5 μg dN6 primers and Superscript-II reverse transcriptase (Invitrogen) following the recommendation of the manufacturer. The gene fragments encoding the variable domain up to the CH2 domain were amplified with the specific primers CALLOOl (5'- GTCCTGGCTCTCTTCTAC AAGG-3') (SEQ ID NO: 102) and CALL002 (5'- GGTACGTGCTGTTGAACTGTTCC-3') (SEQ ID NO: 103) annealing respectively at the leader sequence and within the CH2 exon of the Heavy chain gene of all dromedary IgGs. The 600 bp fragment (VHH-hinge and CH2 but lacking the CHl exon) was eluted from an agarose gel after separation from the 900 bp fragment (VH- CHl-hinge-CH2 exons). Since all VHHs belong to the same family (family III) they can be amplified with one additional PCR with nested primers annealing at the framework-1 (A6E: 5'-GATGTGCAGCTGCAGGAGTCTGGRGGAGG-S') (SEQ ID NO: 104), or SM17: (5'-

CCAGCCGGCCATGGCTSAKGTGCAGCTGGTGGAGTCTGG-S') (SEQ ID NO: 105) and framework-4 (5'-

GGACTAGTGCGGCCGCTGG AG ACGGTG ACCTGGGT-3 ' ) (SEQ ID NO: 106) regions. These primers introduce the restriction sites Pstl, Ncol and NotI, respectively (underlined). The final PCR fragments were ligated, after restriction enzyme digestion, into the phagemid vector pHEN4 [33] cut with the same enzymes. For both libraries (PstI/NotI and NcoI/NotI), ligated material was transformed in E.coli electro competent cells (TGl) and plated on Petri dishes containing LB-agar supplemented with 100 μg ampicillin/ml and 0.1% glucose. The colonies were scraped from the plates, washed and stored at -80°C in LB medium supplemented with glycerol (50% final concentration) until further use.

Selection of Aahll- and Aahl' specific Nanobodies (or binders) The cloned VHH repertoire was expressed on phage after infecting the library with M13K07 helper phages. Recombinant phages were prepared by poly- ethylene glycol (PEG)/NaCl precipitation. Phage particles carrying antigen specific Nbs were enriched by three consecutive rounds of in vitro selection on Aahll or Aahl' (10 μg/well) coated on microtitre plates (Nunc, Maxisorp). Residual protein binding sites on the plates were blocked at room temperature for 2 h with 1% (w/v) casein in phosphate-buffered saline (PBS). In the subsequent step, on average 101 1 phage particles per well were incubated for Ih. After several washing steps with PBS, bound phage particles were eluted for 10 minutes with 100 μl freshly prepared triethylamine (70 μl triethylamine (99%, Sigma) in 5 ml water, pH 10.0). The solution of the eluted virions was transferred to a tube containing 100 μl of 1.0 M Tris-HCl (pH 7.4) and used to infect exponentially growing E.coli TGl cells. An aliquot was properly diluted and plated on LB- Agar with ampicillin. The enrichment of phage particles carrying an antigen-specific VHH was assessed by comparing the number of virions eluted from wells coated with respectively Aahll or Aahl' versus non-Aahll-coated or non-Aahl'- coated wells. After the second and third round of panning, individual colonies were picked and infected with helper phages. The next day, the phage particles released in the culture medium were screened by phage ELISA on microtitre plates coated with 100 μl of AahII or AahF (1 μg/ml). The antigen specificity of the VHH for each clone scoring positive in phage ELISA was confirmed in a direct ELISA with the periplasmic extract from fresh cultures of the original clone induced with ImM Isopropyl-b-Dthiogalactopyranoside (IPTG).

Bivalent and bispecific Nanobody construction and purification Genes encoding toxin binders were amplified with primers MH (5'- CATGCCATGGGAGCTTTGGGAGCTTTGGAGCTGGGGGTCTTCGCTGTGGTG CGCTGAGGAGACGGTGACCTGGGT-3') (SEQ ID NO: 107) and A4short (5'- CATGCCATGATCCGCGGCCCAGCCGGCCATGGCTGATGTGCAGCTGGTGG AGTCT-3') (SEQ ID NO: 108) to introduce an Ncol restriction enzyme site (underlined) at both extremities of the amplified fragment which are take into consideration for specific primer designs. The MH primer also added the hinge sequence of llama IgG2c (N-terminus-Ala-His-His-Ser-Glu-Asp-Pro-Ser-Ser-Lys- Ala-Pro-Lys-Ala-Pro-Met-Ala-C-terminus; SEQ ID NO: 109) to the 3' extremity of the cDNA. The PCR amplified product was then purified with Qiaquick PCR purification kit (Qiagen) and digested overnight with Ncol. The digested fragment was purified again using the same PCR purification protocol. Meanwhile, the recombinant pHENό plasmid containing the same cDNA was digested for several hours with Ncol enzyme and treated with alkaline phosphatase. Finally, the pHENό recombinant plasmid and the PCR fragment were mixed and ligated with T4 ligase. The ligated product was used to transform E. coli WK6 electrocompetent cells. Clones containing the bivalent construct were screened by PCR using FP and RP universal primers. Clones that gave rise to an amplification product of -1000 bp, had their cloned insert sequenced (ABI prism 677 Applied Biosystem), and after expression, the encoded protein was tested for possible binding to specific toxin by ELISA. To obtain larger amounts of the tandemly-linked bivalent molecule, the expression and purification conditions of the monovalent molecule were followed. Construction and purification of Aahl'-Nanobody-human Fc

The gene for human IgGl 'hinge-CH2-CH3' region was introduced in the pCI vector behind the dromedary secretion leader signal and the VHH gene encoding the Aahl'-specific Nanobody. This chimeric plasmid construct was transfected in mouse NSO cells. Subsequent selection of individual clones was performed by ELISA. The supernatant was loaded onto Aahl '-coated wells. Bound Nanobody-Fc was detected using an anti-human IgGl alkaline phosphatase conjugate (Sigma). Clones encoding a chimeric VHH-Fc that scored positive in ELISA against Aahl' were expanded and grown in RPMI medium supplemented with 1% IgG- depleted foetal calf serum. Supernatant containing Nanobody-human Fc was loaded onto a protein- A column, and unbound proteins were washed-out with 100 mM NaCl, 10 mM Tris-HCl (pH 7.5). Pure chimeric Nanobody-human Fc was eluted with 10 mM glycine-HCl (pH 2.5), and dialyzed against PBS. To test the glycosilation of the H-chain, the MW was checked by SDS-PAGE of the reduced sample before or after treatment with N-glycosidase (Roche).

Construction and purification of humanized VHH domains (binders): NbBcIIlO and NbBcII22

Two particular nanobodies referred as NbAahI22 and NbAahIIlO with sub-nanomolar affinity are able to neutralize fully 7LD50 corresponding toxins in mice. The humanizations of these nanobodies have been performed by generation of chimeric Nb constructs by three successive PCR. Both fragments were purified from gel and used in a splincing by overlapping extension (SOE) PCR. Hence, the different CDR-H loops from a donor Nb were transferred to the framework of the recipient NbBcII lO useful as potential candidate. Fragments were digested with Ncol/ BstEII and ligated into the pHEN6 expression vector. Plasmid constructs were transformed into WK6 E.coli cells. Screening was performed by PCR, enzymatic digestion, periplasmic extract ELISA and sequencing. The chimeras format have been tested for their thermodynamic stability, affinity and neutralizing capacity.

Serum stability of the monovalent and the bivalent Nanobodies One μg of monomeric Nanobody and 2 μg of the corresponding tandem linked bivalent format (both at 1 mg/ml) were mixed separately in 200 μl human serum. After different time periods of incubations at 37°C ranging from 0-49 hours, 100 μl was tested in an ELISA on microtiter plates coated with tested toxin. The residual antigen-binding was measured with Nickel-HRP conjugate (ExpressDetector™) to detect the presence of the His-tag at the mono- and bivalent recombinant protein bound to the tested toxin on the microtiter plate. For the chimeric HCAb 7 μg were incubated in 200 μl serum, and to evaluate the residual antigen- binding capacity a mouse anti human Fc monoclonal as detecting agent and a goat anti mouse IgG conjugated to HRP were used.

Thermo stability of monovalent and tandemly linked bivalent nanobodies CD measurements were performed with a JASCO J-715 spectropolarimeter in the far UV (205-250 nm region). The purified protein was prepared in 50 mM sodium phosphate (pH 7.0) at 1.166 mg/ml. A volume of 300 μl was added to the cuvette with a 0.1 cm cell pathlength, and heated from 35°C to 95°C at a rate of 1 °C/min. The fluorescence intensity at 205 nm was recorded as a function of temperature. Data were acquired with a reading frequency of 1/2Os^"1, a 1 s integration time and a 2 nm bandwith. Data analysis was performed assuming a two- state unfolding mechanism.

VHH sequence analysis

The VHH nucleotide sequence of each clone that scored positive in both ELISA's (phage ELISA and direct ELISA) was determined on an automated DNA sequencer (ABI prism 3100 genetic analyzer, Applied Biosystem). The nucleotide sequence was translated into its amino acid sequence. Expression and purification of Nanobodies The VHH genes were subcloned into pHEN6 expression vector using restriction enzymes Ncol or Pstl and BstEII [33]. The plasmid constructs were transformed into the nonsuppressor strain of E. coli WK6 cells. Production of recombinant Nbs was performed in shaker flasks. Freshly transformed cells were grown at 37°C in Terrific Broth, supplemented with 0.1% glucose and 100 μg ampicillin/ml until an OD at 600 nm of 0.6 - 0.9 was reached. The Nb expression was then induced by addition of 1 mM IPTG and fermentation for 16 h at 28⁰C. After pelleting the cells, the periplasmic proteins were extracted by osmotic shock [33]. The recombinant protein was purified from the periplasmic extract by two chromatographic steps: an IMAC (Immobilized Metal Affinity Chromatography) on a His-select (Sigma) column followed by a gel filtration on a 16/60 column Superdex- 75 (Pharmacia). The Nb containing fraction was concentrated on Vivaspin concentrator (VivaScience, Sartorius) with a molecular mass cut-off of 5 kDa. The purity of the protein was checked on an SDS polyacrylamide gel (12%) under reducing and non-reducing conditions and stained with Coomassie blue. The final yield of the purified Nb was determined from UV absorption at 280 tun using the theoretical extinction coefficient and molecular weight, as calculated from the amino acid content of the Nb clone. Solid-phase ELISAs

Maxisorp 96-well plates (Nunc) were coated overnight at 4°C with lOOμl of toxin at 1 μg/ml in phosphate-buffer saline (PBS). Residual protein binding sites in the wells were blocked with 1% (w/v) Casein in PBS for lh30 h at 37°C. The PBS buffer supplemented with 1 % Tween-20 was used for washing steps. Serum IgG ELISA. After incubation with the sera, diluted at 1/5000, the antigen-bound dromedary IgG was detected with a rabbit anti-camel IgG serum. A goat anti-rabbit IgG-horse radish peroxidase conjugate (Sigma) was used as secondary antibody reagent with o-Phenylenediamine dihydrochloride (OPD) as substrate. The colour development was stopped with 50μl 0.2N H₂SO₄ after 10 to 15 min and the intensity was measured at 492 run.

Phage-ELISA. For Phage ELISA, the dromedary IgG was substituted by phage particles. After washings, antigen bound phages were detected using an anti-M13 peroxidase conjugate and staining with peroxydase substrate. The colour was measured at 405nm. Determination of affinity and epitope complementation groups

Binding kinetics between Nb and toxin were assessed by surface plasmon resonance on biosensor instruments (BIAcore-3000 and BIAcore TlOO). The AahII or Aahl' antigen (5 μg/ml in 10 mM acetate buffer, pH 4.6) was immobilized on a CM5 sensor chip using TVhydroxysuccinimide-N-ethyl-N'-(dimethylaminopropyl)- carbodiimide chemistry until 200 resonance units were immobilized. All measurements were performed at a flow rate of 30 μL/min in HBS buffer (10 mM Hepes pH 7.5, 150 mM NaCl, 3.5 mM EDTA and 0.005% Tween-20). Various concentrations of the purified Nb were applied on the chip to determine the association and dissociation rate constants, /.^"on, and /.^"off (using the BIA-evaluation software version 4.1 from BIAcore), from which the equilibrium binding constant, KD, was calculated. After each measurement, the residual Nbs were cleared from the toxin on the surface with a 10 mM Glycine/HCl (pH 1.5) solution. To determine the epitope complementation group(s) of the Nbs, it was injected an excess of a first Nb, after saturating the toxin on the sensor chip with this Nb and reaching equilibrium in binding, it was applied a second Nb in presence of this first Nb. No signal increase will be observed if the second Nb binds to an epitope overlapping that of the first Nb. In contrast, if the signal-increase upon addition of the second Nb equals the signal obtained with this Nb alone (in absence of the first Nb) then evidently, these Nbs bind to non-overlapping epitopes.

Toxicity and neutralization assays of AahII or Aahl' Experiments on mice were carried out in accordance with the European Community Council Directive (86/609/EEC) for experimental animal care, and all procedures met with the approval of the Institutional Research Board of the Pasteur Institute of Tunis. Groups of four healthy, 8 weeks old Swiss mice (20+/-2g) were used to determine the LD50 of the AahII or Aahl' toxin. A volume of 5 or 200 μl of 0.1% BSA, 0.15 M NaCl, containing increasing amounts of AahG50, AahII or Aahl' toxins were individually injected by intracerebroventricular (i.c.v.) or subcutaneous (s.c.) routes, respectively. Control animals received only PBS. The LD50 values were measured according to Bouhaouala-Zahar et al. [34]. The survivors within each cohort of 4 mice that received the same amount of toxin were counted, 24 h after injection. The average of survivors was assessed from two independent experiments, and the LD50 was expressed in ng AahII or Aahl', injected in a mouse whereby 2 out of the 4 mice survived. To measure the toxin neutralizing capacity of the Nanobody, it was employed the standard method to monitor the serotherapeutic batches of equine (Fab)'2 whereby the antibodies (or Nbs in the present case) were premixed at a variable molar excess with an increasing amount of AahII, Aahl' or AahG50.

As a control for toxicity and neutralization assays of AahII, animals received a mixture of 2 LD50 of Aahl' with a 4 molar excess of a purified Nb against Aahl' (NbAahF 22) that does not cross react with Aahll. After a pre-incubation (90 min at 37°C), a constant volume of 5 μl or 200 μl of the mixture was i.c.v. or s.c. injected, respectively, in each mouse (cohorts of 4 mice).

The number of mice surviving the treatment after a delay of 24 h was monitored. The neutralizing capacity (NC) was calculated according to the guidelines proposed by the World Health Organization (1981):

NC (LD50/mg Nb) = (LD50 in presence of Nb - LD50 in absence of Nb) /mg of Nb

EXAMPLE 2: IDENTIFICATION OF POTENT NANOBODIES TO NEUTRALIZE Aahll

Aahll antigen preparation

The AahG50 fraction of scorpion venom obtained after Sephadex G-

50 chromatography was fractionated by Mono-S chromatography. The fraction, highly enriched for Aahll as demonstrated by antigen-specific rabbit antibodies, was injected into an HPLC C8 column. The HPLC analytical run showed a symmetric peak, and the protein within this fraction was characterized as Aahll by its MW of 7249.40 Da on MALDI spectrometry. The median lethal dose (LD50) of the purified Aahll was monitored in 20 g Swiss male mice after i.c.v. and s.c. injection. For i.c.v. administration the Aahll fraction has an LD50 of 3 ng; after s.c. injection LD50 values of 250 ng are recorded. The high toxicity of the Aahll at low doses confirms the high degree of purity of the Aahll preparation.

Humoral immune response elicited in the dromedary

To generate a large panel of high-affinity Aahll-specific Nbs, it was started from a hyperimmunized dromedary. The immunization protocol, reported previously with an AahG50 venom mixture [31], has been adapted and the immunization was started with an Aahll enriched fraction (Mono-S), followed by several boosts with C8-HPLC purified toxin (100 μg). The titre of the dromedary serum (at 1/5000 dilution) during the immunization was followed by ELISA and an antigen-specific response of high titre was induced after 5 weeks. This titre continued to increase after boosts with the pure Aahll. The Aahll-specific immune response was present in both, the conventional antibody and in the HCAb fractions exactly as previously demonstrated [31 ]. Construction of an 'immune' VHH library

The lymphocytes were isolated from anti-coagulated blood of the immunized dromedary. Total RNA was extracted from these lymphocytes and cDNA was synthesized using Superscript-II Reverse Transcriptase (see Experimental procedures). The cDNA was used as template to amplify the gene regions coding for the VHH. After proper restriction enzyme digestion (Pstl and Notl, or Ncol and Notl), the amplified fragments were ligated into the pHEN4 phagemid vector [32] between the PeIB leader signal and the gene III. The ligated material was transformed in E.coli TGl cells to arrive at two 'immune' VHH libraries containing 2x108 and 1.6x108 individual colonies (for Ncol/Notl and Pstl/Notl libraries, respectively). A PCR on 20, randomly chosen, clones from each library to amplify the cloned insert indicated that more than 85% of the clones of the library contained a phagemid with an insert of appropriate size for a VHH.

Enrichment for phage particles carrying Aahll-specific Nanobodies

The VHH repertoire of both libraries (a mixture of equal aliquots containing a representative fraction of each library) was expressed on virions after infection of the bacteria with M13K07 helper phages. Selection of phage particles expressing an Aahll-specific Nb was performed in wells of microtitre plates coated with purified AahII toxin. The enrichment of phage particles carrying an Aahll- specific binder, during the consecutive rounds of panning was followed by a phage- ELISA [32]. After three consecutive rounds of selection on solid-phase coated antigen, a clear enrichment for phage particles carrying Aahll-specific binders was observed. Identification of AahII specific Nanobodies

Forty eight individual colonies were randomly picked after the 2nd and 3rd round of biopanning, infected with helper phages to produce virions that were screened for their capacity to bind to AahII in a phage-ELISA. Thirty seven out of the 48 colonies were shown to produce phage particles carrying an anti-Aahll specific Nb. An RFLP with Hinfl restriction enzyme digestion on the PCR amplified fragment of the cloned Nb genes of these 48 clones yielded about 10 different patterns. The AahII specificity of the Nbs for a selection of 29 clones out of the 37 that scored positive in phage-ELISA was confirmed by culturing the clones, and inducing the expression of soluble Nb into the bacterial periplasm. The proteins from the periplasmic extracts were tested for binding to AahII using a direct ELISA. Twenty eight out of the 29 periplasmic extracts were confirmed to contain Aahll-specifϊc Nbs. All these extracts yield a strong signal except for clones NbAahllOl, 23, 30 and 33 that gave a signal of only about 2x the background signal (background signal is that obtained without immobilising the AahII antigen in the wells). These clones and the one that scored negative (NbAahII47) in this periplasmic extract ELISA were not further analyzed, except for their nucleotide and amino acid sequence. Sequence analysis of selected Nanobodies

The Nb insert of all 37 clones were sequenced and the deduced amino acid sequences are aligned (Figure 1) and numbered according to ImMunoGeneTics numbering (IMGT) [36]. Some of the clones carry a few nucleotide substitutions that encode the same amino acids (silent mutations), these are only shown once in Figure 1. The sequences are grouped in 9 distinct clusters (named A to I) according to their CDR3 sequence homology. Such sequence homology among the binders within a cluster most likely reflects an independent B-cell lineage that was at the origin of these Nbs. Seven of these lineages (A to G) possess the characteristic VHH hallmark amino acids in their framework-2 region with an Arg or Cys at position 50 [27, 37]. As expected for dromedary VHHs, binders from 4 out of the 7 clusters possess an extra pair of Cys apart from the conventional Cys23 and Cys 104. For clones of clusters A, D, and G, one extra Cys is located in the CDR3 and one in the CDRl , whereas clones of cluster C have a Cys in the CDR3 and one at position 50 (i.e. within framework-2). Remarkably, the sequences of one cluster (cluster E) contain only one single extra Cys (in the CDRl). Two clusters (cluster H consisting of one isolated sequence, and cluster I of 14 sequences) contain the VH hallmark amino acids in the framework-2 region (e.g. Leu at position 50). However, the presence of an Arg at position 1 18 instead of the conserved Trp shows that the domains of cluster I will fail to associate with a VL domain [37]. Therefore, also these clones are derived from HCAbs and do not reflect the VH domain that pairs with a VL in a classical antibody. These clones do not have an extra pair of Cys residues in their CDRs (Figure 1). The presence of the single clone NbAahII38 of cluster H is puzzling. It does possess the conserved Trpl 18 in combination with the VH hallmark amino acids in framework-2. Therefore this binder might correspond to a VH domain derived from a conventional antibody, although, during the cloning of the library all efforts were undertaken to avoid such a contamination. The average length of the CDR3 of clones among all clusters is 16 amino acids. The clones that give only a low signal in the periplasmic extract ELISA have the same CDR3 and form one cluster (cluster B, Figure 1). The single clone that fails to yield a signal in the periplasmic extract ELISA (cluster A) has the longest CDR3 of all (24 amino acids). For some clusters, all nucleotide sequences of the individual clones encode an identical amino acid sequence, or one with only a few amino acid differences in their framework regions. In contrast, for other clusters, the sequences contain several amino acid differences, some of them even have a CDRl or CDR2 that differs in length with other sequences from the same cluster. For example, in cluster F, clone NbAahII45 and NbAahII 18 (or NbAahII08) have a CDRl of 9 and 8 amino acids, respectively. This difference could have been provoked by a gene conversion mechanism during affinity maturation [38] or through a cross-over artefact during PCR amplification. Similarly, cluster I members have 7, 8 or 9 amino acids in their CDR2. These different CDR2 lengths within one cluster are also linked to single amino acid substitutions within the CDR3, which makes the argument in favour of gene conversion for their origin stronger, and that of PCR artefact weaker. To investigate the correlation between antigen-specificity (epitope), affinity and AahII toxin neutralizing capacity, 13 clones distributed over all clusters (except cluster A, as the Nbs in the periplasmic extracts of these clones gave poor signal to AahII in an ELISA experiment) were selected for protein expression, and further characterization. Production and purification of soluble monomeric anti-Aahll

Nanobodies

The 13 Nbs were recloned in the expression vector pHEN6, using the restriction enzymes Pstl (or Ncol) and BstEII, and transformed in WK6 competent cells as described previously by Conrath et al. [33]. The anti-Aahll Nbs are expressed with a carboxy-terminal His6 tag to facilitate purification and detection. In addition, the presence of the pelB leader signal transports the recombinant Nb to the periplasmic compartment of E.coli. From the crude periplasmic fraction, the recombinant Nbs are purified to homogeneity by two successive chromatographic steps. After a metal affinity chromatography on a His-select column, the eluted fraction is further purified by gel filtration chromatography on Superdex-75. The Nbs elute in a symmetrical peak, with the notable exception of NbAahIIlO gel filtration profile that shows a minor shoulder corresponding to a maximum of 5% of the material that elutes faster than the monomeric molecules. This corresponds to homodimeric Nbs that dimerized through the Cys from the CDRl (confirmed by running the sample on SDS gels under reducing and non-reducing conditions). Surprisingly, the NbAahII38 with the classical VH sequence imprint has a longer retention time on the gel filtration column than predicted from its MW. It argues for a sticky behaviour and a non-specific interaction between the gel-matrix and this particular Nb. The purity of the proteins within the major elution peak was investigated by Coomassie stained SDS polyacrylamide gels under reducing and non- reducing conditions. For all purified Nbs, only a single protein band at the expected molecular weight of 14,000-16,000 Da is observed. The average yield of each purified Nb varies between 0.60 and 10.00 mg/1 of bacterial culture (grown in shake flasks) and is clone dependent (Table 1 below).

Table 1 : Production yield of recombinant Nbs and their antigen interaction parameters. The production yield of purified Nb per liter of culture is given (last column). The association and dissociation rate constants (kon and fo>fr ) to immobilised AahII were measured by surface plasmon resonance and used to calculate the equilibrium dissociation constant (KD). The name of the actual Nb and the cluster to which it belongs are also given. NbAahII-01 of cluster B is not given as it fails to associate with the AahII toxin immobilized on the chip of the biosensor.

Table 1

Cluster Nanobody K_0n (M-¹S^"1) Koff Cs^"1) K_D (nM) Production yield (mg/I)

-C- NbAahII36 2.04x10⁶ 1.37xlO^"J 0.67 5.20

-D- NbAahII17 6.12xlO^b 3.76x10^"' 0.61 0.60

-E- NbAahIIlO 1.14x10° 5.69x10^"" 0.49 3.50

-F- NbAahlllδ 8.95xlO⁶ 1.60x10^"' 0.18 12.40 NbAahII08 1.19xlO⁷ 1.38xlO^"3 0.12 1.58

-G- NbAahII05 2.77x10° 1.86x10^"' 0.67 10.00 NbAahII39 6.4IxIO⁶ 1.05xl0^"3 0.16 9.30

-H- NbAahII38 1.02x10⁴ 7.8IxIO^"4 76.00 1.44

NbAahII12 5.63x10" 1.60x10^"' 0.28 2.70

-I- NbAahII19 5.46x10⁶ 1.3OxIO^"2 2.39 0.66 NbAahII20 8.18xlO⁶ 2.13xlO^"3 0.26 4.00 NbAahII41 4.33xlO⁶ l.OOxlO^"2 2.31 1.00

Affinity measurements of AahII binders

The affinity and specificity of AahII toxin recognition by the Nbs was measured using surface plasmon resonance. Values for the association and dissociation kinetic rate constants are in the range of 104 M-Is-I to 107 M-Is-I and 5x10-4 s-1 to 10-2 s-1, respectively (Table 1). An example of the affinity sensogram is shown for NbAahIIlO (Figure 2 A). The corresponding equilibrium constants (KD) calculated for the individual binders ranged from 120 pM to 76 nM, and among these 12 clones, 9 Nbs display low KD values of sub nM (Table 1). The two clones from cluster F (NbAahII08 and NbAahll l δ) with one amino acid substitution in their framework-3 region but otherwise sharing an identical sequence in their CDRs possess identical kinetic binding data for AahII. In contrast, the clones NbAahII05 and NbAahII39 from cluster G that differ by 1 amino acid in their CDR2 and 4 out of 12 amino acids in their CDR3, vary four-fold in their KD values. Likewise, the 4 clones originating from the cluster I having slightly different CDR3 sequences have KD values differing by a factor 10. Interestingly, among the clones of cluster I, that with the longer CDR2 (NbAahII20) shows the highest affinity for AahII.

Epitope complementation groups of AahII binders To group the Nbs from the different sequence clusters according to their targeted epitope, the AahII was coupled toxin covalently on the biosensor chip and after saturating the toxin with one Nb, the possible binding of a second Nb (Figure 2 B) was monitored. This epitope mapping was conducted for multiple pairs of Nbs in both directions, i.e. clone NbX followed by binding of clone NbY and vice versa. The interpretation is sometimes complicated by the faster Ic_0H- rate and/or slower Ic_0n of one clone versus that of the other. In addition some binders might induce a conformational shift on a distantly located area of the antigen so that subsequent binding of another Nb is hindered, although both Nbs recognize widely separated epitopes. The data of the epitope complementation grouping are schematically represented in Figure 3. First, it seems that all clones set within the same cluster (based on the CDR3 sequence length and homology) bind to the same epitope complementation group. Consequently, the CDR3 sequence that was used to define the clusters determines which epitope is targeted. Secondly, the antigen binding sites of clones from clusters D and H (NbAahII17 and NbAahII38, respectively) relative to those of the other clones are identical. Their epitopes are therefore expected to have an extensive overlap. Thirdly, the interaction site of the cluster D, G and H binders overlaps with that of clones from clusters F and I, although the two latter binders have clearly an independent epitope. Fourthly, NbAahII36 representing cluster C binds to an independent site relative to the shared epitope of the Nbs from clusters D and H, although all these Nbs fail to recognize the antigen that is in complex with either NbAahII05 or NbAahlll δ. Finally, clone NbAahIIlO that was tested against binders of all other clusters attaches to a unique epitope as all other Nbs could bind equally well to the toxin, independent of the prior presence or absence of NbAahII lO (Figure 2 B). Neutralizing capacity of anti-Aahll Nanobodies The systemic symptoms in mice resulting from the toxic doses of pure AahII or the AahG50 fraction are identical. Following injections, mice develop common symptoms of intoxication, i.e. irritability, jumpiness, agitation, mastication, tachy and bradypnea, although these symptoms can vary in intensity or in time of occurrence according to the injected dose. These symptoms are seriously reduced and/or disappear when the injected toxins are pre-incubated with neutralizing doses of Nbs. The neutralizing capacity of the AahII specific binders from each cluster was first evaluated by an i.c.v. injection of mice with a mixture of a particular Nb with the AahII toxin. This standard experimental approach has previously been shown to be effective to test various antitoxin antibodies in a reliable manner consuming a minimal amount of material [22-24, 31 , 34]. It is a particularly useful test for toxins of which the purification involves a lengthy and tedious procedure.

Neutralization of AahII toxin by Nbs tested by i.c.v. injection. Depending on the actual Nb, the neutralizing activity varies widely (Table 2 below). For example, the Nbs from clusters C, D and H at a 2-fold molar excess neutralize weakly (or not at all) the toxicity of the AahII. In contrast, each one of the five binders from sequence clusters F, G and I neutralizes at least in 50% of the mice -3-4 LD50 of AahII when tested at a 2-4 molar excess of Nb. This corresponds to a neutralizing capacity (NC) for these Nbs at a 4-molar excess to toxin of 27,000 LD50/mg Nb (= 400 LD50/nmol Nb). Remarkably, under similar conditions, the NbAahIIlO of cluster E neutralizes even higher toxic doses of the AahII toxin. Indeed, for a toxin/Nb ratio of 1/4, approximately 168 ng of NbAahIIlO neutralizes 100% the toxicity of 7 LD50 (21 ng) AahII and 240 ng of this Nb neutralizes in 50% of the mice the toxicity of 10 LD50 (30 ng) of AahII. Therefore, according to the NC definition given in the experimental procedures, 1 mg of NbAahIIlO corresponds to approximately 37,500 LD50 or 125 μg of AahII {i.e. 555 LD50 AahII per nmol of NbAahIIlO) (Table 2). The same Nb, at a molar ratio of only 1/1 to the AahII toxin, manages to neutralize for 100% the toxicity of 4 LD50 and 50 % of mice survived with 5 LD50 (Table 2). This data corresponds to a NC value of 133,000 LD50/mg Nb or 2,000 LD50/nmol Nb. As could be expected, a Nb directed against Aahl' (NbAahI'22) that does not cross-react with the AahII in ELISA, fails to neutralize the AahII toxin.

Table 2. In vivo AahII neutralization assay by Nbs from all clusters after i.c.v. administration. In vivo neutralization was assessed by pre-incubating (90 min at 37⁰C) AahII with Nbs at variable molar ratio's as indicated, before i.c.v. injection of various LD50 in Swiss mice. One LD50 of pure AahII was determined at 3 ng / mouse. The Neutralizing Capacity was calculated as NC - (LD50 in presence of Nb - LD50 in absence of Nb) per mg or nmole of Nb, and the data used for this calculation are in bold lettering. The Nbs from cluster C, D and H are barely neutralizing, Nbs from cluster F, G and I are neutralizing to an intermediate extend, and NbAahIIlO of cluster E is the strongest neutralizing antibody. The data in the grey frame corresponds to the maximal amount of LD50 that could be neutralized in 2 out of 4 mice with this Nb at a 4-molar excess to toxin. Table 2

Specific Molar i.c.v. Aahll Nanobody Survivors/ NC = (LD₅₀ with

Nanobody ratio LD_5O dose ng/mouse Injected mice Nb-LD₅₀ without

(cluster) Aahll: /mouse Nb)/Nb

Nb LD₅₀/mg or

LD₅₀/πmol

NbAahII36 (C) 1:2 2 24 1/4

1:2 3 36 0/4 ND

NbAahII17 (D) 1:2 2 24 0/4 ND

NbAahII38 (H) 1:2 2 24 1/4

1:2 3 36 0/4 ND

NbAahII18 (F) 1:2 3 36 1/4

1:2 4 48 0/4

1:4 2 48 2/4

1:4 3 72 2/4 27000 LDjo/mg

400 LDso/nmol

NbAahII05 (G) 1:2 2 24 4/4

1:2 3 36 4/4

1:2 4 48 1/4

1:4 3 72 4/4 ND

NbAahIKO (I) 1:2 2 24 3/4

1:2 3 36 2/3

1:2 4 48 2/4

1:4 3 72 2/4 27000 LD₅₀/mg

400 LD₅₀/nmol

NbAahII19 (I) 1:4 2 48 4/4

1:4 3 72 4/4

1:4 4 96 1/4 ND

NbAahII12 (I) 1:4 2 48 3/4

1:4 3 72 2/4 27000 LDjo/mg

1:4 4 96 0/4 400 LDso/nmol

NbAahIIlO (E) 1:1 4 24 4/4

1:1 5 30 2/4 133000 LDso/mg

1:2 4 96 4/4 2000 LDso/nmol

1:2 5 60 3/4

1:2 6 72 3/4

1:4 2 48 4/4

1:4 3 72 4/4

1:4 4 48 4/4

1:4 5 120 4/4

1:4 6 144 4/4

1:4 7 168 4/4

1:4 8 192 3/4

1:4 9 216 2/4

1:4 10 240 2/4 37000 LD₅₀/mg

555 LDso/nmol

ND : Not Determined Neutralization of AahII toxin by NbAahIIlO tested by s.c. injection. Because of its high NC, the protective activity of NbAahIIlO was further tested after s.c. administering of AahII mixed with this Nb. Taking into account that the LD50 of AahII, s.c. injected in 20 g Swiss mouse is 250 ng (instead of 3 ng for i.c.v. injection), the neutralizing activity of the NbAahIIlO was evaluated at a fixed 4-fold molar excess to 1.5 LD50 (375 ng), 2 LD50 (500 ng), and 3 LD50 (750 ng) of AahII. 3 μg, 4 μg and 6 μg, respectively, of the purified NbAahIIlO were able to neutralize completely the lethal effect of the purified AahII toxin (Table 3). Table 3. In vivo AahII or AahG50 neutralization by NbAahIIlO. The in vivo neutralization was assessed by pre-incubating (90 min at 37°C) the NbAahIIlO at a 4-molar excess to AahII toxin and subcutaneous injection 1.5, 2 or 3 LD50's in mice (top part). The LD50 of pure AahII via this route of injection was determined at 250 ng / mouse. For the experiment shown in the middle part of the table the AahG50 fraction (as opposed to pure AahII toxin) was mixed with a tenfold molar excess of NbAahIIlO before s.c. injection into Swiss mice. The LD50 of AahG50 via this route of injection was determined at 10 μg / mouse. For the lower part of the table 1.5 or 2 LD50 of AahG50 was mixed with a tenfold molar excess of NbAahIIlO and after a pre-incubation of 90 min at 37°C injected the mixture i.c.v. in Swiss mice. One LD50 of AahG50 was determined at 52ng / mouse.

Table 3:

Quantity NbAahIIlO in ng / mouse

(LD50 of AahII /mouse) Molar Ratio Quantity Surviving mice s.c. Ag / Nb (μg / mouse)

375 ng (1.5 LD50) 1 :4 3 4/4 500 ng (2 LD50) 1 :4 4 4/4 750 ng (3 LD50) 1 :4 6 4/4

(LD50 of AahG50 / mouse) s.c.

15 000 ng (1.5 LD50) 1 :10 300 2/4

(LD50 of AahG50 / mouse) i.c.v.

78 ng (1.5 LD50) 1 :10 1.56 4/4 104 ng (2 LD50) 1 : 10 2.80 2/4

Neutralization of AahlIG50 toxic fraction by NbAahIIlO tested by i.c.v. injection. Finally, the neutralization activity of the NbAahIIlO on the AahG50 was assessed. The heterogeneous AahG50 fraction contains different types of toxins (e.g. Na+, K+, Cl channel blockers), which makes it difficult to measure the exact concentration of protein within this fraction. It was considered that an OD at 280 nm of 1 corresponds to 0.54 mg/ml (the same extinction coefficient as for the pure AahII toxin), and found that the AahG50 has an LD50 of 52 ng for i.c.v. injection and of 10 μg for s.c. injection. Because of the presence of various antigenically distinct molecules in the AahG50 it is impossible to neutralize all toxins with a monoclonal binder such as a Nb. However, the neutralization of the AahII alone - the most abundant and toxic molecule - within the AahG50 fraction by NbAahII lO increases the number of survivors. For the i.c.v. injection of 1.5 LD50 (about 78 ng) or 2 LD50 (about 104 ng) AahG50 mixed with an estimated 10-fold excess of NbAahIIlO over the AahII within AahG50, it was noticed that 4/4 and 2/4 mice survived, respectively (Table 3). For an s.c. injection that mimics more closely a scorpion sting, an injection of 15 μg AahG50 (1.5 LD50) kills all mice (4/4), whereas 2/4 mice survived when the same amount of AahG50 was mixed with an estimated 10-fold excess of NbAahIIlO (Table 3). Conclusion

This study reports the identification of a panel of Nbs with a high affinity and neutralizing capacity for AahII scorpion toxin. To obtain these toxin- neutralizing Nbs a dromedary was immunized with AahII toxin. From the blood of the hyperimmune animal, the VHH repertoire of its HCAbs was cloned in a phage display vector and selected multiple Aahll-specific Nbs by phage display panning.

In a previous study, Meddeb-Mouelhi et al. [31] used a heterogeneous AahG50 toxic fraction to immunize the dromedaries and showed that the HCAb isotypes from this immune serum neutralized the scorpion toxin. The AahG50 fraction is still a complex mixture of several proteinaceous toxins of which the AahII is the most toxic molecule [5-6]. Therefore, it was preferred to use an enriched-Aahll preparation to start the immunization and a highly purified AahII toxin for subsequent boosts to elicit a strong and specific immune response against this antigen. A large immune VHH libraries of 1.6 x 108 and 2x108 individual clones was constructed by phage display [27, 32-33], and a panel of Aahll-specific binders (Nanobodies) of high affinity was selected. With the exception of NbAahII38, having an affinity of 76 nM, all other Nanobodies (binders) had a single digit nM or even sub-nM equilibrium dissociation constant. These AahII binders could be categorized in 9 clusters according to the amino acid sequence homology within their CDR3. All members within a particular cluster recognize a topologically linked epitope on the AahII toxin as was demonstrated by BIAcore experiments whereas the sequence variability within the CDRl and CDR2 apparently leads to minor variations in antigen affinity. This confirms that the CDR3 sequence within the Nbs determines the CDR3 loop structure and thus the epitope specificity [39]. Furthermore, it has been reported that the CDR3 of Nbs constitute 60-80% of the antigen contacting area [40], and that a synthetic oligopeptide derived from the CDR3 (peptibody) maintains a measurable antigen affinity and specificity [41]. Hence, peptidomimetics of the CDR3 of one of these Nbs might be designed in the future to arrive at low molecular weight scorpion toxin-neutralizing drugs. In cases were several binders are available to toxic molecules, there is a strong tendency to favour that of highest affinity to develop the most potent toxin neutralizing therapeutics. However, this study clearly illustrates a lack of correlation between the NC and the binding properties (Ic_0n, k_ofr rate constants or KD, the equilibrium constant). It appears that the targeted epitope is far more crucial for optimal toxin neutralization. The Nbs of the epitope clusters C, D and H, possessing most favourable binding characteristics to the AahII toxin perform poor in the AahII toxin neutralization as a twofold molar excess of Nb failed to protect the mice from 2-3 LD50 when i.c.v. administered. In contrast, NbAahII-18, -05, -12, -19 and -20 representatives of clusters F, G and I tested at a 2-4 times molar excess, exhibited an intermediate AahII NC corresponding to ca 27,000 LD50 per mg Nb (or ca 400 LD50 per nmol Nb). It is reassuring to find out that the epitopes of these binders with intermediate NC form a potential continuous stretch when mapped on the toxin surface (Figure 3). More interestingly, it was demonstrated that NbAahII 10 (cluster E) that interacts with a topologically distinct epitope on the AahII possesses the highest NC (ca 37,000 LD50 per mg Nb when tested at a molar AahII/Nb ratio of 1/4, or ca 133,000 LD50/mg Nb when evaluated at an equimolar ratio), despite binders of the other clusters exhibited a faster Ic_0n rate or a higher affinity (lower KD). Since the NbAahII 10 is the binder (Nanobody) with lowest k^ rate (i.e. slowest dissociation of the AahII::Nb complex) it can be argued that this kinetic parameter is critical for the NC. However, a clear correlation between low k_off and the better NC value is absent in the set of Nbs. For example, the strong-neutralizing NbAahII 10 and the non- neutralizing NbAahII38 have an equivalent k_off rate (5.69 x 10-4 and 7.81 x 10-4 s-1, respectively). Likewise, the intermediate-neutralizing binders of cluster F, G and I, and the non-neutralizing binders of clusters C and D have all k_off rates in the same range (1.60 x 10-3 to 3.76 x 10-3 s-1). It was even noticed that, among the binders within cluster I associating to the same epitope, the NbAahII 19 with the fastest k_Ofτ rate performed best in the neutralization assay.

The conclusion that the affinity and binding kinetics are less important parameters for toxin neutralization also explains why the stable Nb (Tm = 60°C) directed against the Aahl' (see Examples 2 below) with moderate weak equilibrium dissociation constant of 58 nM reaches a NC of 17,800 LD50/mg Nb. However, it was realized that this conclusion is at first sight contradictory to the results of the experiments by Juarez-Gonzalez et al. [42]. These authors showed that an affinity maturation of a recombinant scFv (derived form a monoclonal antibody neutralizing scorpion toxin Cn2) from 1 nM to 75 pM (and all binding to the same epitope) has a beneficial effect on the mice survival ratio (protected /injected) when tested at a 10 fold molar excess to 1 LD50 of Cn2 toxin. However, in this study it was also shown that the affinity-matured variant has a higher stability than the original scFv, and it was demonstrated subsequently that the stabilization of the scFv by changing its format into a Fab improves its in vivo toxin neutralization capacity [26],

It is clear that the NbAaMIlO, which targets a unique epitope on Aahll, is the best Nb to combat the lethal effect of this toxic molecule, probably because its epitope overlaps maximally with the Aahll site that interacts with the synaptosomal voltage dependent sodium ion channel. In addition, it might be argued that the surface of the synaptosome receptor site of the scorpion toxin has a higher complementarity to the unique antigen-binding site architecture of Nbs. Indeed, it has been repeatedly documented that Nbs often adopt a unique paratope shape that targets epitopes that are less antigenic for conventional antibodies [27, 40, 43].

According to the i.c.v. experiments, a mixture of 240 ng of NbAahIIlO with 10 LD50 of Aahll (i.e. a 4 molar excess of Nb) neutralizes the lethality of this pure scorpion toxin in 50% of the mice tested. By standard calculations [34], this figure corresponds to an NC of 37,500 LD50 of AahII/mg Nb (or 560 LD50 per nmol Nb). Although the NC values calculated according to this approach are convenient to evaluate batch-to-batch variability of antiserum preparations they cannot be used to compare the therapeutic potency of different neutralizing antibody fragment formats (i e. scFv, Fab, Nb) or when tested on different toxins, alternative injection route (i.p., i.c.v., s.c), or at a different ratio of toxin to antibody. Nevertheless, the complete neutralization of 7 LD50 with a 4 molar excess of Nb, or even 4LD50 at equimolar amount of Nb upon i.c.v. injection remains impressive. A similar potent neutralization of Aahll was noticed with this Nb after s.c. administration of the mixture, which is more relevant as it mimics more closely a scorpion sting. Here, 3 LD50 of the toxin were entirely neutralized with a 4 molar excess of the Nb in all mice. These values demonstrate the potency of NbAahIIlO and it is clear that the neutralization of such high LD50 doses at this low molar excess of antibody were never reached so far with other scorpion toxins, or monoclonal antibody fragments in a scFv or Fab format [22-26, 42].

EXEMPLE 2: IDENTIFICATION OF VHH DOMAINS, BIVALENT VHH DOMAINS AND CHIMERIC HEAVY CHAIN-ONLY ANTIBODIES WITH HIGH NEUTRALIZING EFFICACY FOR AahP

The insert of 13 clones encoding molecules able to bind AaW (VHH domains) with original sequences and deduced amino acid sequences are aligned in Figure 4. The sequences form 1 1 distinct clusters according to their CDR3 indicating that these binders were originated from different B cell lineages A3, B23, and DlO have practically the same CDR3 and a distinguishable CDRl . These clones derived most likely from the same B-cell lineage. The majority of the sequences possess the characteristic VHH hallmark amino acids except Dl 9 that has VH hallmark. All clones have an extra pair of Cys apart from the conventional Cys23 and CyslO4. For clones DlO, C 13, B23, B2 and A3 one extra Cys was located in CDR3 and one in the CDRl . Clone Al 7 contain three Cys in their CDR3 and one in the framework 2 instead of R45 VHH hallmark. Wl 18 present in VHH sequences was substituted by R. This shows that the domains of this VHH fail to associate with a VL domain.

The average length of CDR3 of clones is 18-19 amino acids. The differences in CDRl in length and in amino acids for the clones A3, B23, and DlO could have been provoked by a gene conversion mechanism during affinity maturation.

To investigated the correlation between antigen specificity, affinity and Aahl' neutralizing capacity, the proteins of these 13 clones were expressed for further characterisation.

Hence, these 13 VHH domains were recloned in the expression vector pHEN6, using the restriction enzyme Pstl (or Ncol) and BstEII, and transformed in WK6 competent cells. The anti-Aahl' binders (VHH domains or Nanobodies) were expressed with a Carboxy-terminal His₆ tag to facilitate purification and detection.

Surprisingly, the NbAaWC 13 had on the gel filtration column a retention time that was much later than predicted from MW. It argues for a sticky behaviour and a non-specific interaction between the gel-matrix and this Nb. The purity of the proteins within the major elution peak was checked by coomassie stained SDS polyacrylamide gels. The average yield of each purified Nanobody varied between 0.4 to 10 mg per litre of bacterial culture and was clone dependent.

Affinity measurements of binders:

The affinity and specificity of Aahl' toxin recognition by the Nanobodies was measured using surface plasmon resonances. Results are shown in Table 4 below. Values for the association and dissociation kinetic rates were in the

10 range of 10⁵ M-Is-I to 10⁶ M-Is-I and 10^"4 S-I to 10^"3 s-1 respectively. The Kd value calculated for individual binders ranged from 15.8 10^'9 to 8.62 10^'" M. The two Nb; A3 and F 12 display remarkably low Kd values in the subnanomolar range.

Table 4:

15 Neutralizing capacity of Aahl' VHH domains (also named Aahl'

Nanobodies or Aahl' binders):

Neutralizing capacity of purified Aahl' specific VHH domains was evaluated by injecting mice via i.c.v. route with a mixture of the VHH domain with the Aahl' toxin. This experimental approach has previously been shown to be 20 effective to test various anti-toxin antibodies in reliable manner consuming a minimal amount of material. This is particularly useful in the case of isolated toxins, of which the purification is lengthy and tedious task.

Depending of the actual Nb, the neutralizing capacity (NC) could vary widely. For example B23, A2, C 14 and Bl are not able to neutralise the toxicity 25 of Aahl' with 2 fold molar excess. A3 and Al 7 showed a good neutralising capacity. These two binder have NC that exceed the previous VHH domain selected (NbAahI22).

Remarkably, NbAahI'F12 neutralise 100% the toxicity of 100 LD50 (1.4 microgram) of Aahl' at molar ratio 1 :2 toxin:antibody. This neutralising capacity is exceptionally strong and exceeds all previous antibody constructs that neutralise scorpion toxins. Results are shown in Table 5 below.

Table 5:

Nanobody A3 B2 A17 D19 B23 B l C14 A2 C13 DIO Dl I A19 FI2

Ratio

1 2 1 2 1 2 1 2 I 2 1 2 1 2 1 2 1 2 1 2 I 2 I 2 1 2 I I (Ab-Tox)

4 LD50 4/4 4/4 4/4 4/4 0/4 0/4 0/4 0/4 2/4 3/4 2/4 4 /4 4/4

₅LD5O 1/3 4/4

6 LD50 4/4 4/4 4/4 0/4 2/4 0/4 2/4 -

7 LD50 4/4 1/4 -

8LD5O 4/4 3/4 4/4 1/4 4/4

10LD50 4/4 2/4 3/4

11LD50 1/4

12LD50 2/4 4/4

I3LD50 0/4

20LD50 4/4

23LD50 4/4

26LD50 4/4

29LD₅O 4/4

33LD50 4/4

37LDS0 4/4

47LD50 4/4

60 LD₅O 4/4

70 LD50 4/4

75 LD50 4/4

80 LD50 4/4

90 LD50 4/4

95 LD50 4/4 2/4

100LD50 4/4 2/4

The particular amino acid sequence of NbAahI'22 reveals clearly within high degree of identity with human VH sequences of family III, a VHH sequence imprints. In addition, the first amino acid of the framework-4 (position 1 18) that is invariably a Trp throughout all VHs because it is an important anchoring site for the VL association, has been substituted in this clone to Valine. Figure 5 represents the nucleotide and deduced amino acid sequences of NbAahI'22, according to the IMGT amino acid numbering.

The two bivalent constructs generated from this Nanobody by first linking two NbAahI'22 genes by a spacer of 17 codons and secondly by reconstituting a chimeric HCAb with the NbAahI'22 and the human 'hinge-CH2-CH3'. Thus it has been obtained three NbAahI'22 derived materials: the monomeric Nanobody of MW -14.000; the tandem linked Nanobody of MW -29000 and the humanized chimeric HCAb of MW -78500 (excluding the carbohydrate content in the CH2 domain).

The NbAahl'22 that lacks any linker and the tandem linked bivalent derivative with a natural antibody hinge used as spacer do not aggregate. Moreover, the thermal stability of NbAahl'22 and its derived bivalent form were high (Tm of 60 and 63°C respectively), an indication of a robust entity, although Nbs usually reach even higher thermostability. The monomeric NbAahl'22 and derived constructs were stable at 37°C in human serum for at least 49 hours as checked in ELISA. Longer incubation periods in serum were not tested as proteins of these MW would be totally cleared from the blood via the kidneys after this period of time.

This experimental data on expression levels and stability with the NbAahI'22-based materials confirms the data reported by Conrath et al. [33] and are encouraging to development a Nb based anti-scorpion toxin immunotherapy.

The reconstruction of a chimeric HCAb with NbAahl'22 linked to the human hinge-Fc has been prepared. Its MW is well above the renal clearance cut off and therefore this construct is expected to circulate for longer times in the blood.

Furthermore, it might yield an interesting comparison with the current Fab'2 serotherapies in terms of biodistribution and toxin capturing capacity as both molecules have approximately the same MW, and are both bivalent. The difference lies in the presence of a human Fc in the chimeric HCAb. It was shown that a functional Fc was properly attached as the chimeric molecule was recognized by protein A, and contained the expected N-glycosylation.

The K_D value of the monomeric NbAahl'22 was 55.8 nM. The kinetic binding parameters of the tandem linked bivalent material to the Aahl' toxin were identical to the parental monomeric Nanobody. In contrast, the chimeric HCAb construct showed a net improvement in its k_ofr compared to that of the monomeric

NbAahl'22, leading to a significant 50-fold increased in apparent affinity. Thus, the bivalency of the antigen-binding domain within the HCAb resulted in avidity effects, possibly because the paratopes are sufficiently spaced and optimally oriented for maximal antigen-binding capacity. Thermostability of the monomeric and dimeric Aahl' Nanobodies

The purified Nb and the tandem linked bivalent Nb was brought in sodium phosphate buffer and used to determine the thermal stability (Figure 6). The monomeric Nb has a T_m value of 60⁰C, whereas the bivalent format denatured at slightly higher temperature (63°C). The protein stability is well estimated from calculated T_m values, and apparently, the monomeric Nb and its tandem linked bivalent construct seem to have a robust behaviour.

Serum stability of the monovalent, bivalent and chimeric anti- Aahl' Nanobodies constructs

The monovalent Nb, the tandem linked bivalent Nb, and the Nanobody-human Fc chimeric HCAbs were tested in ELISA for their antigen binding after a variable incubation period in human serum, at 37°C. The antigen-binding capacity of samples incubated in serum for up to 49h, reported as % of residual activity, was indistinguishable from those obtained with freshly purified Nanobodies (Figure 7). This indicates that the Nb and its derivatives remain stable, i.e. active, in normal serum over a time period required for an effective in vivo toxin neutralization.

As negative control, it was tested an irrelevant VHH that not cross react with Aahl' and raised against AahII toxin belonging to another antigenic group than Aahl'.

In vivo neutralizing assay

Obviously, the most critical property of the anti-scorpion toxin antibody is its neutralizing capacity. Initially it was compared the three different

Nanobody formats in their efficacy to neutralize 3LD₅₀ of the Aahl' toxin upon s.c. injection in mice. This amount of toxin was guided from the scorpion toxin neutralizing experiments with a recombinant scFv that, after an in vitro affinity maturation step and at a 10-fold molar excess to Cn2 scorpion toxin, neutralized 1

LD₅₀ of the toxin (Juarez-Gonzalez et al. [42]). These data indicates that all the Aahl'

Nanobody-based materials in a 2/1 molar ratio to toxin, neutralized entirely 3LDs₀ of the Aahl'. By extrapolation, it was calculated that under these experimental settings, at least 70, 200 and 400 LD₅₀ of Aahl' were completely neutralised per mg of chimeric HCAb, tandem linked bivalent Nanobody and monomelic Nanobody, respectively.

An i.e. v. injection pathway of the toxin/antibody was also performed. As before, the ratio of antibody to toxin was kept at a minimum molar excess of 2/1. A 100% reversal of the lethal activity of AahT (i.e. all mice survived) was achieved with 7, 10 and 7 LD₅₀ when mixed with the chimeric HCAb, the dimeric NbAahI'22 or the monomeric Nanobody, respectively. Expressed differently, this corresponds to a total neutralising capacity of 2978, 8928 and 17857 LD₅₀ per mg corresponding to 238, 258 and 267 LDs₀ per nmole of HCAb, dimeric and monomeric Nanobody, respectively. These figures largely surpass the scorpion toxin neutralizing efficacy of scFv-derived fragments reported in the literature that do not reach 200 LD₅₀ per nmole [23, 24 et 42] and are obtained at a lower antibody to toxin molar ratio. At this stage, the exact reason for the excellent neutralisation performance of these Nanobody-constructs remains debatable, although the exact Aahl' epitope recognised by this particular NbAahI'22 antibody fragment might be an essential element.

Conclusion

From the data on Aahl' it can be anticipated that the generation of a tandemly linked bivalent version of the NbAahIIlO might have a higher neutralization capacity against its cognate antigen. Likewise, linking two Aahll-specific Nbs directed against two nonoverlapping epitopes and properly spaced into a single construct might create a chelating effect leading to an extremely potent Aahll- protective agent. It is evident that Nbs, with their small footprint of their paratope [48] on the scorpion toxin antigen and behaving strictly monomeric, are the best choice to design such biparatopic chelators. Alternatively, two Nbs directed against two antigenically different toxic molecules of the scorpion venom (e.g. Aahl' and Aahll) could also be tethered into a similar bispecific construct that might be more versatile to treat scorpion envenoming. Indeed, Devaux et al [6] indicated that the Aah scorpion venom collected from different geographic regions contained a large polymorphism in these immunologically non-cross-reactive toxins. The simultaneous capturing of Aahl' and Aahll by a bispecific 'humanized' Nb construct might therefore be the most appropriate next generation therapeutic to combat the scorpion envenoming. Evidently, for human use it is advisable to 'humanize' the Nbs that after all originated from dromedary antibodies. A generic humanization procedure was recently proposed [47].

REFERENCES 1. Dehesa-Davila, M. and Possani, L. D. (1994) Toxicon 32, 1015-1018.

2. Goyffon, M., et al (1982) Toxicon 20, 337-344.

3. Bouaziz M., et al. (2008) Toxicon 52, 918-926.

4. Jover, E., et al. (1980) Biochem. Biophys. Res. Commun. 95, 1607-1614.

5. Martin-Eauclaire, M. F. and Couraud, F. (1995) Scorpions neurotoxins. Effects and Mechanisms. In Handbook of Neurotoxicity, pp. 683 716, Chang, L. W. & Dyer,

R.S. (Eds), Dekker Inc, New York.

6. Devaux, C, et al. (2004) Toxicon 43, 661-669.

7. Miranda, F., et al. (1970) Eur. J. Biochem .16, 279-291.

8. Rochat, H., et al. (1979) Scorpion toxins: chemistry and mode of action, in Adv. Cytopharmacol. 3, 325-330.

9. Fontecilla-Camps, J. C, et al. (1988) Proc. Natl. Sci. acad. USA. 85, 7443-7447.

10. Housset, D., et al. (1994) J. MoI. Biol. 238, 88-90.

1 1. El Ayeb, M., et al (1983) MoI. Immunol. 20, 697-708.

12. El Ayeb, M., et al. (1986) Biochemistry 25, 6671-6678. 13. Granier, C, et al (1989) MoI. Immunol. 26, 503-513.

14. Chippaux, J.P. and Goyffon, M. (1998) Toxicon. 36, 823- 846.

15. Pepin-Cavotta, S., et al. (1996) Toxicol. App. Pharmacol. 141, 272-277.

16. Ismail, M., (1994) Toxicon 32, 1019-1026.

17. Abroug, F., et al (1999) Lancet. 354, 906-909. 18. Rezende, N.A., et al (1998) Toxicon 36, 1507-1513.

19. Bawaskar, H.S. and Bawaskar P.H. (2007) J. Assoc.Physicians India 55, 14-21.

20. Sofer, S., et al. (1994). J.Pediatr. 124, 973-978.

21. Bahraoui, E., et al. (1988) J. Immunol. 141 , 214-220.

22. Mousli, M., et al (1999) FEBS Lett. 452, 183-188. 23. Devaux, C, et al (2001) Eur. J. Biochem. 268, 694-702.

24. Juste, M., et al. (2007) Cell. MoI. Life Sci. 64, 206-218.

25. Riano-Umbarila L., et al (2005) FEBS J. 272, 2591-2601. 26. Quintero-Hernandez, V., et al. (2007). Mol.Immunol. 44, 1307-1315.

27. Nguyen, V. K., et al (2001) Adv. Immunol. 79, 261-296.

28. Cortez-Retamozo, V., et al (2004) J. Biol. Chem. 279. 1256-1261.

29. Saerens, D., et al. (2005) Anal. Chem. 77, 7547-7555. 30. Baral, T. N., et al. (2006) Med. Sci. 22, 914-916.

31. Meddeb-Mouelhi, F., et al. (2003) Toxicon 42, 785-791.

32. Arbabi-Ghahroudi, M., et al. (1997) FEBS Lett. 414, 521-526.

33. Conrath, K. E., et al. (2001). Antimicrob. Agents and Chemother. 45, 2807-2812.

34. Bouhaouala-Zahar, B., et al. (1996) Eur. J. Biochem. 238, 653-660. 35. Hmila H., et al. (2008) MoI. Immunol. 45, 3847-3856.

36. Lefranc, M. P., et al. (2003) Dev. Comp. Immunol. 27, 55-77.

37. Muyldermans, S., et al. (1994) Protein Engineering. 7, 1 129-1 135.

38. Schiaffella, E., et al (1999) J. Immunol. 162, 3984-3995.

39. De Genst, E., et al. (2005) J. Biol. Chem. 280, 141 14-14121. 40. De Genst, E., et al. (2006) Proc. Natl. Acad. Sci. U S A. 103, 4586-4591.

41. Marquardt, A., et al. (2006) Chemistry 12, 1915-1923.

42. Juarez-Gonzalez, V. R., et al. (2005) J. MoI. Biol. 346, 1287- 1297.

43. Desmyter, A., et al (1996) Nature Struct. Biol. 3, 803-81 1.

44. Muyldermans, S. (2001) J Biotechnol. 74, 277-302. 45. Yvonne, E., et al. (2002) Enzyme and Microbial Technology. 30, 273-278.

46. Leon G. J Frenken., et al (2000) Journal of Biotechnology. 78, 1 1-21.

47. Vincke, C, et al. (2009). J. Biol. Chem. 284, 3273-3284.

48. De Genst, E., et al. (2006) Dev. Comp. Immunol. 30, 187-198.

49. Martin, M. F., and Rochat, H. (1984) Toxicon 22, 695-703. 50. Rochat, H., et al. (1972) Eur. J. Biochem. 28, 381-388.

51. Kopeyan, C., et al. (1979) Eur. J. Biochem. 94, 609-615.

52. Darbon, H., et al. (1982) Int. J. Pept. Protein Res. 20, 320-330.

53. El Ayeb M, and Rochat H. (1985) Toxicon. 23, 755-60.

54. De Dianous, S., et al. (1987) Toxicon 25, 731 -741. 55. Delabre, M. L., et al. (1995) Biochemistry 34, 6729-6736.

Claims

1°) An isolated variable domain of a camelid heavy-chain antibody (VHH domain) directed against an Androctonus australis hector (Aah) toxin or the isolated CDR3 region therefrom. 2°) The VHH domain according to claim 1 , characterized in that the

Aah toxin is Aahl' or Aahll.

3°) The VHH domain according to claim 2, characterized in that said VHH domain is directed against Aahl' and is selected from the group consisting of AahI'B2 (SEQ ID NO: 1), Aahl'Bl (SEQ ID NO: 2), AahI'A17 (SEQ ID NO: 3), AahI'A3 (SEQ ID NO: 4), AahI'F12 (SEQ ID NO: 5), AahI'D19 (SEQ ID NO: 6), AahI'DIO (SEQ ID NO: 7), AahI'C14 (SEQ ID NO: 8), AahI'C13 (SEQ ID NO: 9), AahI'B23 (SEQ ID NO: 10), Aahl'Dl l (SEQ ID NO: 1 1), AahI'A2 (SEQ ID NO: 12), AahI'A19 (SEQ ID NO: 13) and NbAahI'22 (SEQ ID NO: 1 1 1).

4°) The VHH domain according to claim 3, characterized in that it is AahI'F12 (SEQ ID NO: 5) or NbAahI'22 (SEQ ID NO: 1 1 1).

5°) The VHH domain according to claim 2, characterized in that said VHH domain is directed against Aahll and is selected from the group consisting of NbAahII47 (SEQ ID NO: 27), NbAahIIlbis (SEQ ID NO: 28), NbAahII23 (SEQ ID NO: 29), NbAahIDO (SEQ ID NO: 30), NbAahII33 (SEQ ID NO: 31), NbAahII15 (SEQ ID NO: 32), NbAahlllό (SEQ ID NO: 33), NbAahII36 (SEQ ID NO: 34), NbAahIB (SEQ ID NO: 35), NbAahllό (SEQ ID NO: 36), NbAahII17 (SEQ ID NO: 37), NbAahIIl l (SEQ ID NO: 38), NbAahII32 (SEQ ID NO: 39), NbAahII lO (SEQ ID NO: 40), NbAahII2bis (SEQ ID NO: 41), NbAahII9 (SEQ ID NO: 42), NbAahII08 (SEQ ID NO: 43), NbAahll lδ (SEQ ID NO: 44), NbAahII45 (SEQ ID NO: 45), NbAahII05 (SEQ ID NO: 46), NbAahII48 (SEQ ID NO: 47), NbAahII39 (SEQ ID NO: 48), NbAahII38 (SEQ ID NO: 49), NbAahII29b (SEQ ID NO: 50), NbAahII37bis (SEQ ID NO: 51), NbAahII41 (SEQ ID NO: 52), NbAahII43 (SEQ ID NO: 53), NbAahII20 (SEQ ID NO: 54), NbAahII21 (SEQ ID NO: 55), NbAahII44 (SEQ ID NO: 56), NbAahII4bis (SEQ ID NO: 57), NbAahII28bis (SEQ ID NO: 58), NbAahII 12 (SEQ ID NO: 59), NbAahII 13 (SEQ ID NO: 60), NbAahII 14bis (SEQ ID NO: 61), NbAahII35 (SEQ ID NO: 62) and NbAahII 19 (SEQ ID NO: 63). 6°) The VHH domain according to claim 5, characterized in that it is NbAahIIlO (SEQ ID NO: 40).

7°) A VHH domain according to anyone of claims 1 to 6, characterised in that it is obtainable by the method comprising the steps of: (a) immunizing a camelid, preferably a dromedary, with a Aah toxin,

(b) isolating peripheral lymphocytes of the immunized camelid, obtaining the total RNA and synthesizing the corresponding cDNAs;

(c) constructing a library of cDNA fragments encoding VHH domains, (d) transcribing the VHH domain-encoding cDNAs obtained in step

(c) to mRNA using PCR, converting the mRNA to phage display format, and selecting the VHH domain by phage display, and

(e) expressing the VHH domain in a vector and, optionally purifying the expressed VHH domain. 8°) An isolated CDR3 region according to claim 1, characterized in that the CDR3 region derived from a VHH domain directed against Aahl' toxin is selected from the group consisting of SEQ ID NO: 14 to 26.

9°) An isolated CDR3 region according to claim 1 , characterized in that the CDR3 region derived from a VHH domain directed against AahII toxin is selected from the group consisting of SEQ ID NO: 64 to 100.

10°) An isolated polypeptide, characterized in that it comprises at least one VHH domain according to anyone of claims 1 to 7 or at least one CDR3 regions according to claims 1 , 8 or 9.

1 1°) An isolated polypeptide according to claim 10, characterized in that it comprises at least two different VHH domains according to anyone of claims 1 to 7 or at least two different CDR3 regions according to claims 1 , 8 or 9.

12°) An isolated polypeptide according to claim 10 or to claim 1 1, characterized in that it comprises a human Fc antibody fragment and at least one VHH domain according to anyone of claims 1 to 7 or at least one CDR3 regions according to claims 1 , 8 or 9. 13°) An isolated polynucleotide, characterized in that it encodes a VHH domain according to anyone of claims 1 to 7, a CDR3 region according to claims 1, 8 or 9 or a polypeptide according to anyone of claims 10 to 12.

14°) A recombinant expression cassette, characterized in that it comprises a polynucleotide according to claims 13, under control of a transcriptional promoter.

15°) A recombinant vector, characterized in that it comprises a polynucleotide according to claim 13 or a recombinant expression cassette according to claim 14. 16°) A host cell, characterized in that it contains a recombinant expression cassette of claim 14 or a recombinant vector of claim 15.

17°) A therapeutic or diagnostic agent comprising a VHH domain according to anyone of claims 1 to 7, a CDR3 region according to claims 1, 8 or 9, or a polypeptide according to anyone of claims 10 to 12. 18°) A diagnostic agent according to claim 17, characterized in that the said VHH domain, CDR3 region or polypeptide is linked, directly or indirectly, covalently or non-covalently to a diagnostic compound.

19°) A method for diagnosing a subject envenomed by Androctonus australis hector, comprising the steps of: a) contacting in vitro or ex vivo an appropriate biological sample from said subject with a VHH domain according to anyone of claims 1 to 7, a CDR3 region according to claims 1 , 8 or 9, a polypeptide according to anyone of claims 10 to 12, or a diagnostic agent according to claims 17 or 18, b) determining the presence or the absence of an Aah toxin in said biological sample, the presence of said Aah toxin indicating that said subject is envenomed by Androctonus australis hector.

20°) A pharmaceutical composition comprising a therapeutic agent according to claim 17, and a pharmaceutically acceptable carrier.

21 °) A VHH domain according to anyone of claims 1 to 7, a CDR3 region according to claims 1 , 8 or 9, a polypeptide according to anyone of claims 10 to

12, a polynucleotide according to claim 13, or a diagnostic agent according to claims 17 or 18, for use in the treatment of envenoming and intoxication by Androctonus australis hector.

22°) A kit for diagnosing or monitoring, in a subject, an envenoming and an intoxication by Androctonus australis hector, comprising at least a VHH domain according to anyone of claims 1 to 7, a CDR3 region according to claims 1 , 8 or 9, a polypeptide according to anyone of claims 10 to 12, a polynucleotide according to claim 13, or a diagnostic agent according to claims 17 or 18.