PHARMACEUTICAL COMPOSITION COMPRISING ANTIBODIES AGAINST A TOXIN

The present invention relates to pharmaceutical compositions comprising antibodies and antibody fragments which bind specifically to a toxin, which are useful in the treatment of poisoning by said toxin. Methods of preparing the antibody and antibody fragment components of the compositions and methods of treatment of toxin poisoning using these compositions form further aspects of the invention.

Of particular interest is the treatment of ricin poisoning. Ricin is a stable toxin which may be extracted from the seeds or beans of the castor oil plant, Ricinus communis. Castor beans are grown agriculturally in many parts of the world, including the Middle and Far

East, for the commercial production of castor oil. The plants are also known to grow wildly in arid locations, including in the United States. Although ricin is approximately 1 ,000-fold less toxic than the botulinum toxins, it is nonetheless toxic and potentially fatal if ingested. It is thought that a fatal dose can be ingested simply by a child chewing just 3 castor beans or by an adult eating 8 seeds. By injection, a fatal dose is as little as between 1 and

3 micrograms per kilograms of body weight. Historically, ricin has been employed in criminal activities and recently it has been considered as potential weapon for extremist and terrorist groups. This is perhaps because of its high availability and the ease and speed with which large quantities can be produced.

The ricin protein itself is a 66 kilodalton (K _d3 ) glycoprotein cytotoxin (toxalbumin) consisting of two polypeptide chains, called the A-chain and the B-chain, which are linked by an easily split disulphide bond. The toxin binds via its B-chain to terminal galactose residues found on the surface of many cells and via both A and B-chains to mannose receptors found on certain populations of cells such as macrophages and hepatic sinusoidal liver endothelial cells. These binding events trigger the endocytic uptake of the toxin into the cell. Once internalised, the protein translocates into the cytosol where the A chain enzymatically inactivates ribosomes, by removing a single adenine residue from the 28S ribosomal subunit. This inactivates the 60S ribosomal subunit disrupts protein synthesis and inhibits protein synthesis resulting in eukaryotic cell death. The catalytic action means that a single ricin molecule has the potential to inactivate many ribosomes and consequently can kill a cell. In addition to causing intracellular toxicity following

internalisation, the A-chain can have direct extra cellular effects on cells, causing vascular leak.

The pathological effects and subsequent clinical signs of ricin intoxication depend on the route of exposure, as this dictates the subsequent tissue distribution of the toxin. For example, animal studies have indicated that the lethal dose of ricin is similar for systemic

(e.g. intravenous or intraperitoneal routes) and inhalational exposures, but is higher for intragastric administration, which reflects poor absorption of the toxin from the gastrointestinal tract. Inhalational exposure produces effects that are mainly confined to the respiratory tract. Following intravenous or intramuscular administration, lesions develop in the spleen, liver and kidneys but the lung is not affected. After oral ingestion the gastrointestinal tract is severely affected, which ultimately may prevent the liver, spleen and kidneys from functioning. In all cases of ricin poisoning, there is a delay between exposure to ricin and the appearance of toxic effects. The length of the delay is dependent on the dose of ricin and the route of exposure but during this delay or lag phase, although there are no visible signs of intoxication, irreversible damage has been caused to cells.

The concept of using animal or human-derived polyclonal antibodies and antidotes or antisera against toxins is well established. Such antibodies, or antitoxins, are able to neutralise toxins and can thus provide passive immunity to individuals when administered either before (prophylactically) or after (therapeutically) exposure to a toxin. Antitoxins can be routinely produced following the immunisation of animals using an inactive form of a toxin (e.g. non-toxic subunit or toxoid) or by low levels of the toxins themselves. Antitoxins produced in animals can be purified and used to treat other animals or humans that have been exposed to the toxin. Whereas it is relatively easy to immunise individuals prophylactically with antibodies and provide protection against toxin exposure, postexposure therapy is more difficult to achieve due to the irreversibility of systemic damage once a toxin has entered a host cell.

Furthermore, there is always a risk when using antitoxins, such as whole IgG or IgM, of producing unwanted side effects in patients. It is possible that animal derived antibodies will be recognised as foreign and causes the host immune system to generate immune responses against the antibody. This can result in immediate adverse effects following a

single exposure (anaphylactoid) or anaphylactic reactions on subsequent antitoxin administration.

Several researchers have proposed the use of antitoxin fragments to reduce the risk of such side effects. These antitoxin fragments, such as Fab', are despeciated since the Fc portion of the antibody is removed. This is thought to reduce the risk of adverse side effects. However it is often the case that a despeciated antibody fragment is less effective than a whole antibody in providing a therapeutic effect when administered as an antitoxin.

Co-pending international patent application, WO 2004/106376 has shown that compositions which comprise either whole antibodies or large binding fragments in combination with small binding fragments can be effective in treating the effects of certain toxins, such as botulinum toxins. This is attributed to the speed at which small antibody fragments are dispersed throughout the body to the sites of action of the toxin.

However, in spite of these well known approaches to antibody therapy, there is no still no effective prophylactic or post-exposure therapy available for the treatment of persons poisoned by certain toxins, such as ricin or staphylococcal enterotoxin B (SEB).

The applicants have now found that a composition which comprises a whole antibody which is specific for the target toxin together with a large binding fragment of such an antibody is particularly effective in the treatment of toxin poisoning and, in particular, for the treatment of ricin intoxication. The applicants have observed that such a composition comprising a combination of a whole antibody and a large binding fragment is significantly more effective than administering either individual component alone. In addition to providing increased levels of protection, compositions of the present invention are particularly advantageous in that significantly lower levels of intoxication are observed in animals, when compared to using either component alone.

According to the present invention, there is provided a pharmaceutical composition comprising an antibody which binds specifically to a target toxin and a large binding fragment of an antibody which binds specifically to said toxin.

The antibody may be any immunoglobulin such as IgG, IgM, IgE, IgA, IgD or IgT or any subclass thereof but in particular is an IgG or IgT. Preferably the antibody is an IgG. If desired the antibody may be humanized using conventional methods, or comprise a chimeric antibody.

The antibodies may be obtained using any conventional means, as is well understood in the art. For example, polyclonal antibodies may be generated by immunisation of an animal (such as rabbit, rat, chicken, goat, horse, sheep, cow etc) with a toxin mimic, the toxin itself or immunogenic subunits or fragments of these to raise antisera, from which antibodies may be purified. The immunization of the animal may utilize an adjuvant as is necessary. In one embodiment both the whole antibody and the large binding fragment are derived from polyclonal antibodies. Conveniently the polyclonal antibodies may be derived from the same or combined batches of antisera.

Alternatively the antibodies may be derived from monoclonal antibodies. Monoclonal antibodies may be obtained by fusing spleen cells from an immunized animal with hybridoma cells, and selecting cells which secrete suitable antibodies. This process is well understood in the art.

The large binding fragment of the composition may be any antibody fragment provided it comprises a significant proportion of the antibody from which it is derived. For instance, it will comprise the entire variable domain, as well as some of a constant region (Fc). In particular, large antibody fragments include F(ab') ₂ or F(ab) ₂ fragments but they may also comprise deletion mutants of an antibody sequence. In particular the large binding fragment is F(ab') ₂ .

Such large binding fragments are also suitably derived from polyclonal or monoclonal antibodies using conventional methods such as enzymatic digestion with enzymes such as pepsin to produce F(ab') ₂ fragments. Alternatively the fragments may be generated using conventional recombinant DNA technology.

The antibodies and the large binding fragments used in the composition of the invention may be derived from the same or different sets or source of the antibody. They may be

specific for the same or different antigens, provided that the antigens are associated with the same toxin.

In particular the antigen is associated with a toxin which is required to be inactivated in a patient. The toxin may be present as a result of exposure to the toxin, For example, toxins, such as botulinum toxin, anthrax toxin or ricin toxin may be inhaled in biological warfare situations or in laboratory accidents or they may be ingested in food containing the toxins. The latter applies also to Staphlococcal enterotoxins, such as SEB, which are typically associated with food poisoning. In particular, however, the toxin is ricin.

Compositions the invention may comprise antibodies and large binding fragments which bind more than one toxin molecule, for example a range of toxins produced by the same microorganism or animal. Thus the antibodies and fragments may be multivalent in nature, or they may be specific for antigens which are common to more than one toxin. Alternatively the compositions may comprise more than one set of antibodies and large binding fragments, each set being specific for a different toxin molecule.

Pharmaceutical composition according to the present invention may further comprise pharmaceutically acceptable carriers or excipients as are well known in the art. These may be solid or liquid carriers depending on the intended mode of administration.

Any desired mode of administration may be used, and this will depend upon factors such as the nature of the toxin being treated, and the nature of the patient. In particular compositions of the invention will be intended for oral, parenteral (especially intravenous) or intranasal administration, or for administration by inhalation or insufflation.

Oral compositions may be in the form of tablets, lozenges, hard or soft capsules, aqueous or oily suspensions, emulsions, dispersible powders or granules, syrups or elixirs. Compositions for parenteral administration will suitably be in the form of a sterile aqueous or oily solution for intravenous, subcutaneous, or intramuscular dosing.

Compositions for intranasal administration or for administration by inhalation or insufflation will suitably comprise a finely divided powder, and inhalable compositions may also be in the form of a liquid aerosol.

Compositions of the invention may comprise other components such as preservative agents, inert diluents, granulating and disintegrating agents, binding agents, lubricating agents, anti-oxidants as well as colouring, sweetening or flavouring agents, depending on the nature of the composition.

Compositions for administration by inhalation may be in the form of a conventional pressurised aerosol arranged to dispense the active ingredient either as an aerosol containing finely divided solid or liquid droplets. Conventional aerosol propellants such as volatile fluorinated hydrocarbons or hydrocarbons may be used and the aerosol device is conveniently arranged to dispense a metered quantity of active ingredient.

The relative amounts of pharmaceutically acceptable carrier to the antibody and large binding fragments in a formulation will vary depending upon factors such as the particular route of administration. Generally however, compositions will comprise from about 1 to about 98 percent by weight of pharmaceutically acceptable carrier, and preferably from 5 to 90 percent by weight of pharmaceutically acceptable carrier.

The size of the dose for therapeutic purposes of a composition of the invention will naturally vary according to the nature and severity of the condition, the age and sex of the animal or patient and the route of administration, according to well known principles of medicine. Generally however, patients are given from 0.5 mg to 75 mg per kg body weight of the antibody and large binding fragment.

Furthermore, pharmaceutical compositions of the invention may also further comprise an adjuvant as is well known in the art.

Compositions of the invention are suitably administered to a patient in need thereof, as soon as possible after exposure to the toxin. In the case of a poisoning incident, such as a snake or scorpion bite, this may be carried out as soon as possible after the incident has occurred. In the case of toxins produced by microorganisms, which have infected a patient and where exposure is not known, the compositions are suitably administered as soon as the exposure event is detected or as soon as symptoms are noted.

The window of opportunity will vary depending upon the particular patient or animal exposed, and the dosage of the toxin.

In a further aspect the invention provides a pharmaceutical composition comprising an antibody which binds specifically to a target toxin and a large binding fragment of an antibody which binds specifically to said toxin, for use in the treatment of poisoning by the toxin.

Such pharmaceutical compositions as described above are particularly useful for the treatment of ricin poisoning or intoxication

Thus, in yet a further aspect, the invention provides the use of a pharmaceutical composition comprising an antibody which binds specifically to a target toxin and a large binding fragment of an antibody which binds specifically to said toxin, in the manufacture of a medicament for the treatment of toxin poisoning.

As described above the composition preferably comprises an IgG antibody and an F(ab') ₂ fragment. A particularly preferred use of such pharmaceutical compositions is in the manufacture of a medicine for the treatment of ricin poisoning.

In yet another aspect the invention provides a method of treating toxin poisoning comprising administering an effective amount of a composition comprising an antibody which binds specifically to the toxin and a large binding fragment of an antibody which binds specifically to the toxin.

The composition may be administered before toxin exposure occurs, for example, as a result of an accidental of deliberate dissemination of the toxin or as a result of a detection event which indicates that the risk of exposure is high. However, the composition is suitably administered as soon as possible after exposure to the toxin has taken place. Repeat administrations may be necessary. The method of the invention is particularly effective in treating the effects of ricin poisoning.

The invention will now be described by way of example with reference to the accompanying drawings in which:

Figure 1 shows IgG levels in sheep serum during and after immunisation using different immunogens and adjuvants. Sheep were immunised subcutaneously with 20μg immunogen in alhydrogel or FIA on weeks 0, 2, 6.12 and 17 and test bled on weeks 0, 2, 6, 12, 14 and 19 analysed for levels of anti-ricin IgG in ELISA.

Figure 2 shows an in-vivo comparison of IgG efficacy (post exposure) in mice. Balb/c mice (6 per group) were injected with IgG (1.3mg) (iv) 2-hrs following ricin (5 x LD50) (ip). The number of survivors in each group is shown for each day following challenge. Controls were administered ricin (intraperintoneally) and PBS (intravenous) only.

Figure 3 shows the post-exposure efficacy of IgG, F(ab')2 and Fab' (administered individually) against ricin challenge. Balb/c mice (6/group) were injected with ricin (ip) (2.5 x LD50) and IgG, F(ab')2 and Fab' (iv) 1 hr later. Number of survivors is shown.

Figure 4 shows a comparison of symptom scores (b) and weight loss (a) for different antitoxin treatments given 2 hour after ricin (2.5xLD50). Balb/c mice (6/group) were administered IgG (2.6mg), F(ab')2 (2mg) , IgG (1.3mg) with + F(ab')2 (1 mg) or F(ab') 2 (1 mg) with + Fab' (1 mg) (i.v.) 2 hours after ricin (2.5 LD50) (i.p.) The mean symptom scores (a) and weight loss (b) for mice in each group are shown. The numbers in parenthesis indicate the number of survivors. Symptom scoring included 0 = no symptoms, healthy mouse, 1 = slight piloerection, 2 = medium piloerection, 3 = severe piloerection, 4 = severe piloerection and decreased mobility, 5 = severe piloerection and total immobility and 6 = death

Figure 5 shows a pharmacokinetic comparison of IgG, F(ab')2 and Fab' in mice. Balb/C mice (4 per group) were injected via the intravenous route with 2mg IgG, F(ab')2 or Fab'. Blood samples were taken at t=1 , 6 and 24 hours. Plasma levels of each antitoxin as determined by ELISA are shown.

Production of anti-ricin Antitoxins

Immunogens used to raise antisera in sheep were the ricin recombinant A-chain (supplied by M Lord, University of Warwick), the purified ricin B-chain (Vector Laboratories) and ricin toxoid. Ricin toxoid was produced by formaldehyde treatment of whole, purified ricin.

Groups of 5 Scottish half-bred sheep were used in this study. Each sheep was immunised subcutaneously with 20μg of immunogen adsorbed to alhydrogel adjuvant (Biofors, Denmark) in phosphate buffered saline (PBS)(20% v/v) on weeks 0, 2, 6, 12 and 17. A total volume of 1ml was administered per dose to each sheep, divided between 4 separate injection sites (two sites in the neck and two axillary sites) to ensure better dispersal. Prior to each dose of immunogen and two weeks after the final boost, test bleeds were taken to enable the monitoring of serum anti-ricin IgG levels. Test bleeds were taken on weeks 0, 2, 6, 12, 14 and 19. Following in vitro and in vivo testing of the IgG produced from three groups of sheep a fourth group was immunised with an emulsion of ricin toxoid in Freunds Incomplete Adjuvant (FIA) (Brenntag, Denmark) (50% v/v) in PBS as described above. Production bleeds were taken on weeks 21 and 24.

Monitoring of anti-ricin IQG concentrations in sheep serum bv ELISA

Serum levels of anti-ricin IgG were monitored in test bleeds taken throughout the immunisation schedule using an enzyme-linked immunosorbant assay (ELISA). Wells of lmmulon 96- well microtiter plates (B. E. Thompson Supplies, Andover, UK) were part- coated with either 5μg/ml ricin toxin (produced in house) or a standard curve of commercial sheep IgG (Sigma-Aldrich, Poole, Dorset, UK) and incubated overnight at 4 ⁰ C. Plates were washed three times with 0.05% Tween-20 in PBS (washing buffer) and blocked with PBS containing 1% BSA (Sigma-Aldrich, Poole, Dorset, UK) (blocking buffer) for 1 hr at 37 ⁰ C. Serum samples, diluted in 1% BSA in 0.05% tween 20 in PBS (dilution buffer) were added to designated wells of the microtitre plates. All standard wells were maintained at an equal volume using dilution buffer. The plates were then incubated for 1 hour at 37 ⁰ C. After washing three times with washing buffer, horse radish peroxidase (HRP) labelled donkey anti-sheep secondary antibody (Sigma-Aldrich, Poole, Dorset, UK) (diluted 1 in 5000 in buffer) was added and the plates incubated 37 ⁰ C for 1 hour. Following washing, ABTS (Sigma-Aldrich, Poole, Dorset, UK) substrate in citrate buffer (citric acid @11.77g/L, Na ₂ HPO ₄ @12.49g/L) produced in-house, was added to each well and incubated for 30 minutes. The absorbance at 414nm was read and concentrations of anti-ricin IgG in serum samples determined from standard curves.

Production of whole IgG

Polyclonal IgG was purified from production bleeds taken on weeks 21 and 24 of the immunisation schedule. A process that has been validated for human use products was

used. This was carried out by the Protein Fractionation Centre based at the Scottish National Blood Transfusion Service (SNBTS) in Edinburgh.

Production of despeciated antitoxin fragments (F(ab')? and Fab') from IqG of sheep immunised with ricin toxoid with FIA

Despeciated antibody fragments were prepared from the IgG produced from sheep immunised with whole ricin toxoid in FIA using papain digestion.

Determination of percentage ricin specific IqG and relative binding affinities of each IqG product

The percentage anti-ricin IgG in each polyclonal IgG product was measured using a modified ELISA. 96 well microtitre plates were part-coated with either 5μg/ml ricin toxin or a commercial sheep IgG standard curve (Sigma-Aldrich, Poole, Dorset, UK). The ELISA was carried out as described above with serial dilutions of the IgG products of known protein concentration added to designated wells of blocked plates in dilution buffer. Dilution buffer only was added to wells of the standard curve. Following development of the ELISA the percentage anti-ricin IgG in each dilution of test antitoxin was determined from comparison of dilutions of samples to the standard curve and the mean value calculated. The binding affinity of the anti-ricin IgG samples was also estimated by determining the concentration at which the binding to the immobilised ricin was half- maximal.

In vitro neutralising activity of anti-ricin IqG in a BPAE cell model of ricin toxicity

Bovine lung endothelium (BPAE) cells (Cell Line no. 86123101) were obtained from the European Collection of Animal Cell Cultures (ECACC), Health Protection Agency at Porton Down, Salisbury and were used to study the in vitro toxicity of ricin. The cells were maintained in DMEM (Sigma-Aldrich, Poole, Dorset, UK) containing 15% (v/v) Fetal Calf Serum (Sigma-Aldrich, Poole, Dorset, UK), 1% (v/v) Penicillin/Streptomycin solution and 1% (v/v) L-Glutamine (Sigma-Aldrich, Poole, Dorset, UK). BPAE cells were grown as previously described using 150cm ² flasks in a humidified atmosphere of 5% CO ₂ /95% at 37 ⁰ C and removed from the flask surface in by incubation with trypsin (0.05% w/v) containing EDTA (0.02% w/v) on achieving 90% confluency. In order to sediment the cells, the suspension was centrifuged at 25Og for 5 min. The cells were resuspended in

the required volume for passaging into new flasks or resuspended in the appropriate volume of medium for seeding into 96 well test plates.

In vitro cytotoxicity studies The concentration of ricin that gave a 70% inhibition of cell growth (IC ₇₀ ) compared to non- ricin treated controls was determined as follows. BPAE cells were seeded into wells of 96-well flat bottomed cell culture plates (B. E. Thompson Supplies, Andover, UK) at a cell density of 5 x 10 ⁴ cells per well. The cells were allowed to adhere to the culture plates for 24 hours and then exposed to different concentrations of ricin in culture medium. 8 replicates of each ricin concentration were used per plate. Cell growth was measured 24 hours later using a gentian violet stain as previously described and the IC ₇₀ calculated.

In order to determine the in vitro neutralising activity of the anti-ricin IgG products, known concentrations of IgG were mixed with the IC ₇₀ of ricin for one hour before adding to adhered BPAE cells. The assay was then carried out as described above. The concentration of each antitoxin required to restore BPAE cell growth to that of non-ricin treated controls (i.e. to totally neutralise an IC ₇₀ dose of ricin) was determined. From this the relative potency of each IgG product could be determined as well as the molar ratio of total and ricin specific IgG required to neutralise ricin calculated.

Protective efficacy of anti-ricin IaG products in mice

Age-matched female Balb/c mice purchased from a designated supplier, (42-49 days old; Charles River Laboratories Ltd, Margate, Kent, UK) were used in all animal studies. The mice were housed in a Home Office designated establishment in rooms maintained at 21 ⁰ C +/- 2 ⁰ C on a 12/12-hour dawn to dusk cycle. Humidity was maintained at 55% +/- 10% with an airflow of 15-18 changes/hour. Mice were kept in polycarbonate shoebox- type cages with steel cage tops and corncob bedding (International Product Supplies, Wellingborough, UK). Mice were fed a standard pelleted Teklad TRM 19% protein irradiated diet (Harlan Teklad, Bicester, UK) and given fresh water daily, ad libitum. All animal experiments adhered to the 1986 Scientific Procedures Act, and were carried out under appropriate ethically approved licenses from the UK Home Office. In order to confirm the results of the in vitro efficacy studies and to enable a decision to be made on which IgG to despeciate to F(ab') ₂ and Fab', a study was carried out to compare the in vivo protective efficacy of the four anti-ricin IgG products in mice. Balb/c mice (6

per group) were randomly assigned to treatment groups and weighed on day 0 of the experiment. Antitoxins (IgG) were administered intravenously 2 hours following the intraperitoneal administration of.5 x LD ₅₀ ricin (5 x the dose which kills 50% of mice). Mice were weighed and checked daily for signs of intoxication and time to death (if applicable) recorded. Untreated groups of age-matched mice were used as controls.

Post-exposure efficacy studies for comparison of IaG. F(ab') _g and Fab' from sheep immunised with ricin toxoid in FIA

A study was carried out to compare the in vivo efficacy of the IgG, F (ab ¹ ) ₂ and Fab' antitoxins. Antitoxins were administered intravenously 1 hr following a 2.5 x LD ₅₀ ricin challenge in mice. Mice were weighed and checked daily for signs of intoxication and time to death (if applicable) recorded.

A second study was carried out to compare the therapeutic efficacy of combinations of antitoxins. Mixtures of F(ab') ₂ with IgG or Fab' were compared to IgG and F(ab') ₂ alone. For this study, doses of antitoxins were used that reflected equal numbers of antigen binding sites in the total antitoxin dose administered. Antitoxins were administered intravenously 2 hours following a 2.5 LD ₅₀ ricin challenge in mice. Mice were weighed and checked daily for signs of intoxication and time to death (if applicable) recorded.

In vivo pharmacokinetic studies using IqG. F(ab% and Fab'

Pharmacokinetic studies were carried out in approximately 6 week old, female Balb/c mice. The levels of IgG, F(ab') ₂ and Fab' in plasma following iv administration at various time points were investigated. 2mg of antitoxin (IgG, F(ab') ₂ or Fab') was administered intravenously to each mouse (n=4). Plasma samples were taken by cardiac puncture at t=10 mins, 1 hr, 6 hr and 24 hrs using halothane anaesthesia and added to lithium heparin tubes (Teklab Ltd, Durham, UK). Samples were gently mixed by inversion and centrifuged at 10,000 rpm for 10 minutes and the supernatant aliquoted into sterile eppendorfs and frozen at -2O ⁰ C

ELISA analysis of anti-ricin IaG. F(ab')? and Fab' concentrations in plasma samples from pharmacokinetic studies.

Plasma levels of anti-ricin IgG, F(ab') ₂ , and Fab' in mouse plasma were analysed using the enzyme-linked immunosorbent assay (ELISA). 96-well microtiter plates were coated

with 5μg/ml ricin toxin and incubated overnight at 4 ⁰ C. Plates were washed three times with washing buffer and blocked with blocking buffer for 1 hour at 37 ⁰ C. Serial dilutions were made of each antitoxin fragment in naive Balb/c plasma and 1 in 10 or 1 in 20 dilutions made in dilution buffer to produce a standard curve of 100μl per well in triplicate. Plasma samples were diluted 1 in 10 or 1 in 20 in dilution buffer and added to designated wells on the microtitre plates (100μl per well) in quadruplet. The plates were sealed and incubated for 1 hour at 37 ⁰ C and then washed three times with washing buffer. HRP-labelled donkey anti-sheep secondary antibody (diluted 1 in 5000 in dilution buffer) was added and the plates sealed and incubated at 37 ⁰ C for 1 hour. Following washing, ABTS substrate was added to each well and incubated for 30 minutes. The absorbance at 414nm was read and the concentrations of anti-ricin IgG, F(ab') ₂ and Fab' in plasma samples determined from standard curves.

Results

Monitoring of anti-ricin IgG concentrations in sheep serum bv ELISA

Serum anti-ricin IgG concentrations were determined in test bleeds taken on weeks 0, 2,

6, 12, 14 and 19. For the three groups of sheep immunised using the alhydrogel adjuvant the highest anti-ricin IgG concentrations were obtained using ricin toxoid as the immunogen. However the levels of anti-ricin IgG were difficult to sustain between immunisations (as shown in Figure 1). Following these results and those of preliminary in vitro and in vivo studies where the three IgG's were compared (results not shown), a further group of sheep was immunised using ricin toxoid in FIA to see whether this improved plasma IgG concentrations. Plasma levels of anti-ricin IgG from the group of sheep immunised with ricin toxoid in FIA steadily increased to a peak of 42mg/L at the final test bleed at week 19 (Figure 1) which was similar to that for ricin toxoid/alhydrogel immunised sheep. Of more importance however was the observation that plasma IgG levels were sustained between immunisations for the ricin toxoid/FIA immunised sheep.

Percentage ricin specific IgG and relative binding affinities of each IQG product The amount of anti-ricin specific IgG as a percentage of the total IgG in each product was determined by ELISA. The results are shown below in Table 1. The IgG made from the plasma of sheep immunised using ricin toxoid and FIA adjuvant had approximately 15-fold

more anti-ricin specific IgG than the other IgG products. The affinity of the ricin specific IgG in each product was determined using a modified ELISA. No significant differences were seen between the IgG products (table 1).

Table 1 - In vitro characterisation of IqG produced from plasma of ricin toxoid, ricin A- chain and ricin B-chain immunised sheep using alhvdroqel adjuvant or FIA. Specific activity of each IgG product was determined using ELISA. Molar ratios for ricin neutralisation were determined by examining what concentration of ricin specific IgG fully neutralised a 7-pmol/L ricin solution in vitro using BPAE cells to assess toxicity. Data are means ± SE; n equals number of experiments. Data comparison using 2-sample t-test; d compared to a, b, c = P < 0.05. P > 0.05 for data comparisons e-h.

In vitro neutralising activity of anti-ricin IaG

The concentration of ricin that gave a 70% reduction in the growth of BPAE cells compared to control untreated cells was 7pM. The concentrations of the different IgG products to fully neutralise this concentration of ricin were determined. From this the number of moles of total IgG required to neutralise one mole of ricin was calculated. The molar ratio of ricin specific IgG required to neutralise one mole of ricin were calculated based on the percentage anti-ricin specific IgG in each product. The results are shown in

table 1. The ricin toxoid/FIA IgG was the most efficacious product requiring 15 fold less product to neutralise ricin than the ricin toxoid/alhydrogel IgG. The ricin toxoid/FIA produced IgG, however, required a similar number of ricin specific molecules of IgG to neutralise ricin as the toxoid/alhydrogel and A-chain alhydrogel IgG. The ricin B- chain/alhydrogel IgG, however, was the least effective antitoxin, requiring approximately 20-fold more molecules of ricin specific IgG to neutralise each ricin molecule (table 1).

Protective efficacy of anti-ricin IqG in mice

The ability of the four IgG products to protect against ricin challenge in mice were compared. Antitoxins (IgG) were administered intravenously 2 hours after a 5 LD ₅₀ ricin challenge. Control mice (ricin only) and those treated with the ricin A-chain/alhydrogel and B-chain/alhydrogel IgG died within 24 hour of ricin administration (figure 2). An extended time to death was seen in mice given ricin toxoid/alhydrogel IgG. 100% survival was seen in mice given the ricin toxoid/FIA IgG although some signs of ricin intoxication were seen. These included weight loss, piloerection, diarrhoea, abdominal pinching and reduced mobility.

In vivo comparison of IgG. F(ab')? and Fab' produced from sheep immunised with ricin toxoid in FIA A preliminary study was carried out to compare the protective efficacy of the IgG, F(ab') ₂ and Fab' antitoxins when administered 1 hr following a 2.5 x LD ₅₀ ricin challenge in mice. Balb/c mice treated with IgG and F(ab') ₂ all survived whereas only 1/6 mice treated with Fab' survived (figure 3). For ethical reasons it was decided not to use Fab' on its own for any further protection studies as it was clearly not protective at the dose used. Although all mice treated with IgG and F(ab') ₂ survived, some signs of ricin intoxication were seen, including piloerection, weight loss, diarrhoea and abdominal pinching. This was more apparent in the F(ab') ₂ treated mice. A scoring system was used to compare the visible signs of intoxication.

In a second study the protective efficacy of combinations of F(ab') ₂ with IgG and F(ab') ₂ with Fab' was compared to IgG and F(ab') ₂ alone (figure 4). This was to see whether combinations of antitoxins (fragments and IgG) are more effective than individual antitoxins. In this study antitoxins were administered intravenously 2 hours following a 2.5 LD ₅₀ dose of ricin. IgG and F(ab') ₂ both gave 100% survival and some weight loss was

seen in both groups. This was most apparent for the F(ab') ₂ antitoxin (figure 4a). Mice given F(ab') ₂ also showed mild signs of intoxication including piloerection and mild pinching of the abdomen which was not evident for the IgG treated mice (Figure 4b). Over the two week period following the ricin challenge, IgG treated mice gained weight and achieved their starting weight whereas F(ab') ₂ treated mice did not. The protection afforded by combinations of F(ab') ₂ with IgG and F(ab') ₂ with Fab' were compared to IgG and F(ab') ₂ alone. The combination of F(ab') ₂ with IgG resulted in the survival of 5/6 mice with initial weight loss similar to that observed with IgG (fig 4a) alone and no other visible signs of intoxication (fig 4b) . However weight gain was more rapid for mice treated with the combination of IgG with F(ab') ₂ than IgG or F(ab') ₂ alone (figure 4a). The combination of F(ab') ₂ and Fab' gave 100% survival although weight loss and other signs of intoxication were similar to F(ab') ₂ administered alone with prolonged weight loss observed (figure 4b). On post-mortem the F(ab') ₂ and Fab' combination group peritoneal abdominal haemorrhage was observed which was not seen in mice given only F(ab') ₂ .

Discussion

This study shows that it is possible to produce antitoxins to ricin that can be used for the post-exposure therapy of ricin intoxication in mice.

Comparison of serum levels of anti-ricin IgG using ricin toxoid, ricin A-chain and ricin B- chain with alhydrogel adjuvant showed that throughout the immunisation schedule the toxoid generally produced the highest levels of anti-ricin IgG. However there was a tendency for the levels of IgG to drop when extended times were used between immunisations (figure 1). The purified IgG products of all three groups of sheep however, contained approximately the same percentage of anti-ricin IgG.

In vitro testing of the three IgG products in the BPAE cytotoxicity assay showed that all three products could neutralise ricin. However, at the molecular level, the IgG from ricin toxoid and A-chain immunised sheep was approximately 20-fold more effective than the B-chain IgG at preventing ricin toxicity. This suggests that in this model, internalisation following binding via the A-chain or direct toxicity of the A-chain in the absence of internalisation, possibly through apoptosis is more important than toxicity mediated through the binding of the B-chain and subsequent toxin internalisation.

Of the three IgG products produced using the alhydrogel adjuvant, the toxoid IgG was the most effective product for the post exposure therapy of ricin in mice with the A-chain and B-chain IgG products being ineffective when administered alone. This suggests that in vivo it is necessary to block the activity of both sub-units of the ricin molecule. These results are in contrast to other toxins consisting of two sub units such as for example botulinum toxin where binding and internalisation occurs solely via a domain on the non toxic heavy chain. With these toxins antibodies against the heavy chain (e.g. produced through vaccination with a subunit containing the binding domain) are effective at preventing toxicity.

Although the ricin toxoid IgG gave the best protection against ricin intoxication out of the three alhydrogel products by extending the time to death in mice, it did not actually prevent death when administered alone. It is possible that this observation is due to a low percentage of anti-ricin IgG being present in the in the final product. We therefore investigated whether using a different adjuvant would increase the percentage of anti-ricin IgG in the product and its protective efficacy. FIA was used as an alternative adjuvant since this adjuvant is likely to be suitable for use in creating a human-use product, made to GMP guidelines. After administration of the FIA together with ricin toxoid, a comparison of serum samples taken during the immunisation period showed a steady increase in anti- ricin IgG levels which were sustained between injections (figure 1). Differences in the final IgG product were apparent with a 15-fold more anti-ricin IgG in the FIA product. This may reflect the period of time between the final test bleed and the production bleeds (4 weeks and 7 weeks) and the tendency for IgG levels to drop when alhydrogel is used as an adjuvant. In contrast, IgG levels are sustained or even increase with time when FIA is used as the adjuvant. A similar tendency was seen when alhydrogel and FIA were used as adjuvants to immunise sheep with botulinum toxoids.

When compared to the three alhydrogel adjuvantised products the IgG product produced from sheep immunised with ricin toxoid with FIA was 15 fold more effective than the ricin toxoid/alhydrogel IgG at neutralising ricin in vitro. This was due to increases in the amount of ricin specific IgG in the product (15 fold) rather than any changes to the affinity of the IgG produced. It was also much more effective in vivo, with 100% survival seen in mice given the ricin toxoid/FIA IgG 2 hours following 5 LD ₅₀ ricin. This compares to the other

three IgGs that at best, gave an extended time to death (as seen with the ricin toxoid/alhydrogel product) indicative of partial protection. Of the four individual IgG products compared, the ricin toxoid/FIA IgG product was the most effective for the therapeutic treatment of ricin poisoning. However some signs of intoxication were still observed.

Immediate adverse effects of administering IgG derived from foreign host species to humans can arise from the triggering of anaphylactoid and anaphylactic responses . Non- human antibody fragments are less likely to cause anaphylaxis if the Fc region of the antibody molecule (which causes the host to recognise them as 'foreign') has been removed. When this process is utilised for animal antibody preparations, it is known as despeciation. However, it has been reported previously that these antibody fragments were less immunogenic in mice (Mayers, C. N., Veall, S., Bedford, R. & Holley, J., 2003. lmmunopharmacology and Immunotoxicology 25 (3), 397-408)

However, when administered intravenously 1 hour following a 2.5 x LD ₅₀ challenge with ricin via the intraperitoneal route, IgG and F(ab') ₂ both independently gave 100% survival whereas Fab' did not give any protection. It is possible that these results indicate that for protection against ricin intoxication it is necessary to sustain a minimum level of antitoxin in the blood and that Fab' leaves the blood so rapidly it may not maintain a high enough level in the blood. The pharmacokinetic properties of small fragments such as Fab' are such that the fragment achieves rapid and wide extravasal distribution. This, however, does not appear increase its effectiveness as an antitoxin against ricin.

Studies have shown that following intravenous and intraperitoneal injection in mice, ricin rapidly binds to cells in the blood, liver and spleen. Extrapolation of pharmacokinetic data to time t = 0 suggests that 50% of the administered ricin has been removed from the plasma within minutes of its administration. There is therefore a very small window within which antitoxins can effectively be administered. This explains why IgG and F(ab') ₂ alone were each unable to prevent all the signs of ricin intoxication. The signs of intoxication seen in the antitoxin treated mice are indicative of incomplete ricin neutralisation.

Surprisingly, however the administration of a composition comprising whole antibody, such as IgG together with a large binding fragment such as F(ab') ₂ was significantly more

effective than either individual component at preventing ricin intoxication and, in particular, ricin-induced weight loss. This shows that a synergistic effect is obtained when administering mixtures of antibody and large antitoxin fragments to treat toxin poisoning. Not only does such a combination increase the quality of the protection afforded when compared to individual antibody components but it also significantly reduces the effects of intoxication. This is particularly surprising since the antitoxin fragments on their own were less effective than IgG. Of particular note was the increased rate at which pre- experimental starting weights were achieved when IgG was administered in combination with F(ab') ₂ compared to IgG alone. It seems likely that for the post-exposure therapy of systemically administered ricin in mice antitoxin needs to be present and maintained in the plasma as well as reaching in the extracellular fluids. The combination of IgG and F(ab') ₂ achieves this whilst significantly reducing the effects of intoxication.