|Home | About | Journals | Submit | Contact Us | Français|
Antitoxins for botulinum neurotoxins (BoNTs) and other toxins are needed that can be produced economically with improved safety and shelf-life properties compared to conventional therapeutics with large-animal antisera. Here we show that protection from BoNT lethality and rapid BoNT clearance through the liver can be elicited in mice by administration of a pool of epitope-tagged small protein binding agents together with a single anti-tag monoclonal antibody (MAb). The protein binding agents used in this study were single-chain Fv domains (scFvs) with high affinity for BoNT serotype A (BoNT/A). The addition of increasing numbers of differently tagged scFvs synergistically increased the level of protection against BoNT/A. It was not necessary that any of the BoNT/A binding agents possess toxin-neutralizing activity. Mice were protected from a dose equivalent to 1,000 to 10,000 50% lethal doses (LD50) of BoNT/A when given three or four different anti-BoNT scFvs, each fused to an E-tag peptide, and an anti-E-tag IgG1 MAb. Toxin protection was enhanced when an scFv contained two copies of the E tag. Pharmacokinetic studies demonstrated that BoNT/A was rapidly cleared from the sera of mice given a pool of anti-BoNT/A scFvs and an anti-tag MAb but not from the sera of mice given scFvs alone or anti-tag MAb alone. The scFv pool and anti-tag MAb protected mice from lethality when administered up to 2 h following exposure of mice to a dose equivalent to 10 LD50 of BoNT/A. These results suggest that it will be possible to rapidly and economically develop and produce therapeutic antitoxins consisting of pools of tagged binding agents that are administered with a single, stockpiled anti-tag MAb.
Microbial toxins are the cause of many serious human diseases, and several of these toxins are listed among the NIAID category A and B priority pathogens. Specifically, botulinum neurotoxin (BoNT) is a category A threat, and ricin, epsilon toxin, Staphylococcus enterotoxin B, and Shiga toxins are category B threat agents. Other microbial toxins, such as those produced by Clostridium difficile, Clostridium tetani, Staphylococcus, Bordetella pertussis, and Corynebacterium diphtheriae, also cause serious human diseases. Toxins produced by some animals and plants, such as insects, spiders, snails, snakes, and jellyfish, also can cause human disease. Currently, most toxin-induced diseases are treated with antitoxins, usually polyclonal antisera produced in animals. These antitoxins are generally produced by immunizing large animals with a chemically inactivated form of the toxin or a nontoxic portion of the toxin that elicits polyclonal antisera which bind the holotoxin and prevent its uptake into cells and/or accelerate its clearance. While effective, antitoxins are often expensive to manufacture and problematic for quality control, and they have a limited shelf life. Furthermore, products derived from serum can cause serum sickness and may have pathogen contamination. There is thus a clear need for improved agents to prevent or treat intoxication.
Botulism is a flaccid paralysis resulting from exposure to BoNT, usually by an oral route (reviewed in reference 24). This toxin is considered among the most dangerous biodefense threats because of its extreme potency, wide availability, relative ease of production, stability, and lack of specific treatment modalities (5). Seven different BoNT serotypes (A to G) are known to exist, and further subtype variants are found within some serotypes. Each BoNT serotype contains a 100-kDa heavy chain, responsible for transcytosis across mucosal membranes and for neuron receptor binding and internalization, and a 50-kDa light chain protease that cleaves one or more SNARE proteins to inactivate neurotransmitter exocytosis. Because antidote therapies do not exist, significant intoxication by BoNT will result in death due to paralysis of the muscles associated with respiration unless constant, intensive, and prolonged supportive care is provided.
Botulism symptoms can be reduced or prevented if proper antitoxins are administered within about a day following exposure to moderate doses. While polyclonal antitoxin products are available to treat human botulism, it is widely accepted that improved antitoxin agents are badly needed (27). One approach showing promise is to replace polyclonal antisera with monoclonal antibodies (MAbs) (1, 2, 4, 9, 17, 20). This approach has been applied successfully in mice, using anti-BoNT MAbs (17). Although a single MAb was weakly protective, pools of two and three different MAbs led to synergistic improvements in efficacy. The requirement for perhaps 21 different MAbs to protect against all seven known BoNT serotypes and problems relating to high MAb development and production costs and limited shelf life are likely to hinder widespread stockpiling of such antitoxin agents. Here we provide evidence that a pool of small tagged antitoxin binding agents administered with a single anti-tag MAb protects mice with the same efficacy as that of polyclonal antitoxin sera. Therapeutic agents of this type should be more rapid to develop and more economical to produce and stockpile than conventional antitoxins.
BoNT/A holotoxin was obtained from Metabiologics Inc. For pharmacokinetic studies, homogeneous botulinum toxin type A (BoNT/A) was isolated from bacterial cultures as previously described (7, 21, 25). The homogeneity of the materials was confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (14). All studies with holotoxin were performed within a biosafety level 2 select-agent laboratory registered with the CDC. Equine polyclonal anti-BoNT/A was generously provided by the CDC.
DNAs encoding the BoNT/A heavy chain carboxyl-terminal receptor binding domain (A-HC) and the light chain protease domain (A-LC) were synthesized with codons optimized for Escherichia coli expression. The A-HC coding DNA, encoding amino acids 861 to 1296 of BoNT/A1 (GenBank accession no. EDT83034), was ligated into the pQE30Xa (Qiagen) expression vector. The A-LC coding DNA, codon optimized for E. coli expression and encoding amino acids 1 to 448 of BoNT/A1, was ligated into the pET14b (Novagen) expression vector. Recombinant expression was induced as recommended by the manufacturers, and the soluble proteins were purified by standard nickel-affinity chromatography.
Six sheep were immunized with BoNT/A antigens, using different immunogens and adjuvants. The sheep (animal 249) that reached the highest neutralizing antibody titer (1 μl of antibody protected mice from the equivalent of 10,000 50% lethal doses [LD50] of BoNT/A1) had been immunized initially with 250 μg of recombinant BoNT/A1 heavy chain carboxyl end (A-HC) in complete Freund's adjuvant, followed by three monthly boosts with 250 μg A-HC in alum-CpG. Subsequently, the sheep received multiple, increasing, approximately weekly doses (0.1, 0.25, 0.5, 0.75, 1.5, and 3 μg) of BoNT/A1 holotoxin (Metabiologics Inc.) in phosphate-buffered saline (PBS). Several months later and prior to tissue harvest, the sheep received another 250-μg boost of A-HC and two additional weekly doses of 2 μg BoNT/A1 holotoxin. Peripheral blood lymphocytes (PBLs) were obtained from blood, and cDNA was produced from PBL mRNA by reverse transcriptase, using random hexamer and oligo(dT) primers as previously described (23).
PCR primers were employed to amplify the VH and VL coding regions, containing the antibody diversity in the sheep, using previously established primer design and methods (16). The VL and VH domains were sequentially cloned into the JSC phage display vector (23) to produce a library with about 7 × 106 phage, with >80% containing inserts with both VH and VL domains. This scFv display library, representing the antibody repertoire of the BoNT/A-immunized sheep, was panned on Immunotubes (Nunc) coated with 1 μg/ml BoNT/A holotoxin, using standard procedures. To enrich for higher-affinity scFvs, subsequent rounds of panning employed reduced amounts of holotoxin coated on the tubes (10 ng/ml), reduced binding times, and more extensive washing. Following three or more panning rounds, colonies expressing soluble anti-BoNT/A scFvs were identified by enzyme-linked immunosorbent assay (ELISA). About 100 positive clones were characterized by DNA fingerprinting using the BstNI and/or HaeIII restriction enzyme and, for some, by DNA sequencing.
DNAs encoding unique scFv clones were transferred to the JSC-his periplasmic expression vector (23) such that each scFv was produced with an epitope tag, or E tag (GAPVPYPDPLEPR), and six histidines at the carboxyl end. In one instance, a second copy of the E tag coding DNA was engineered at the amino-terminal coding end of scFv#7 (scFv#7-E2). JSC-his expression plasmids for each scFv were transfected into the E. coli Rosetta-gami strain (Novagen), and fresh overnight cultures were prepared in LB medium containing ampicillin (50 μg/ml), 2% glucose, and chloramphenicol (34 μg/ml). The cultures were diluted about 10× in 500 ml of LB medium with ampicillin (50 μg/ml) and shaken in a flask until they achieved an A600 of 0.7 to 1. IPTG (isopropyl-β-d-thiogalactopyranoside) was added to 0.5 mM, and the culture was shaken overnight at 30°C. Bacteria were pelleted by centrifugation, resuspended in 10 ml of PPB buffer (PBS with 200 mg/ml sucrose, 1 mM EDTA, 30 mM Tris, pH 8.0), and stored on ice for at least 20 min. Bacteria were again centrifuged, resuspended in 10 ml of 5 mM MgSO4 in PBS, and stored on ice for at least 20 min. Following centrifugation, the MgSO4 supernatant was combined with the PPB supernatant and extensively dialyzed against PBS. Following dialysis, the scFv was purified from the periplasmic extract by use of Ni-nitrilotriacetic acid (Ni-NTA) agarose beads. Before use, about 2 ml of a 1:1 slurry of Ni-NTA beads (Invitrogen) was allowed to settle for 30 min at room temperature, the unbound liquid was removed, and the beads were washed once with 5 ml of PBS-25 mM imidazole. The beads were then added to the dialyzed periplasmic extract and rocked gently at room temperature for 30 to 60 min before being loaded onto a prerinsed filter column. The flowthrough was passed once more through the column, and then the column was washed twice with 5 ml of PBS-50 mM imidazole. One milliliter of elution buffer (PBS, 0.5 M imidazole) was added and allowed to enter the column and sit for 5 min at room temperature before flowing through. This process was performed three times, and the 3 ml of eluate was then pooled. The elution buffer was then replaced by PBS, using concentration and diafiltration on iCon 9K (Pierce) filters or by dialysis. Protein purity was checked and yields quantified by SDS-PAGE with protein controls.
ELISAs were performed to detect scFv recognition of native BoNT/A holotoxin and purified heavy chain (A-HC; Metabiologics) and recombinant BoNT/A light chain (A-LC) protease. The BoNT/A holotoxin (1 μg/ml), A-HC (5 μg/ml), or A-LC (5 μg/ml) was used to coat ELISA plates and incubated with serial dilutions of the purified scFvs. Bound scFvs were detected by a horseradish peroxidase (HRP)-conjugated anti-E-tag MAb (GE Healthcare). Biacore analysis was performed as recommended by the manufacturer. Chips were coated with BoNT/A and surface plasmon resonance measured following the addition of a recombinant scFv, or chips were coated with recombinant scFv and resonance measured upon the addition of soluble BoNT/A. Kd (dissociation constant) values were assessed by surface plasmon resonance.
Neuronal cerebellar granule cells were obtained as previously described (26) and cultured in Dulbecco's modified Eagle's medium (DMEM)/F-12 medium (Gibco-Invitrogen, Carlsbad, CA) supplemented with 50 U/ml penicillin, 50 μg/ml streptomycin, 10% fetal bovine serum (FBS), N2 supplement, and 25 mM KCl. Cells were seeded onto poly-l-lysine-coated six-well plates at a density of 2.2 × 106 cells per well and maintained in a humidified 5% CO2 atmosphere at 37°C. After 24 h, 10 μM cytosine-β-d-arabinoside (Sigma, St. Louis, MO) was added to the culture to inhibit replication of nonneuronal cells. The neurons were maintained by replacement every 7 days with the same freshly prepared medium and were used for assays within 4 weeks. For the assay, 0.1 nM BoNT/A holotoxin was preincubated with scFv preparations or controls for 30 min at room temperature and then added to the culture medium. The cells were incubated for 3 h at 37°C, washed, and harvested into a preweighed Eppendorf screw-cap vial. After cell pelleting, the supernatant was carefully removed and the microcentrifuge tubes weighed. Four microliters of M-PER (Pierce) and 2 μl of SDS-PAGE sample buffer were added for each mg of cell pellet, and samples were boiled for 10 min to inactivate the residual toxin. Proteins were resolved by 15% SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) filters (Millipore) for standard Western blotting. SNAP-25 was detected with rabbit anti-SNAP-25 antibody (1:2,500; Sigma) followed by HRP-conjugated goat anti-rabbit IgG (1:10,000; Amersham), using an ECL kit (GE Healthcare). Band densities for SNAP-25 and its cleaved product were normalized and relative intensities determined using scanning densitometry (Kodak Image 200RT station).
Murine lethality studies were conducted under an approved animal care and use protocol. One day prior to each study, female CD-1 mice (Charles River, Wilmington, MA) were weighed and sorted to reduce intergroup (n = 5) weight variation. BoNT/A1 was dosed on the basis of pg/g of body weight, using the average weight (19 to 20 g) of all mice included within each experimental study. When coadministered with toxin, the scFv preparations or controls, with or without the E-tag MAb, were incubated with BoNT/A1 in PBS containing 0.1% gelatin (Sigma, St. Louis, MO) for 30 min at room temperature prior to intravenous (i.v.) administration in a 200-μl volume/mouse. Control animals received BoNT/A1 alone in PBS containing 0.1% gelatin. Mice were observed 2 to 6 times/day for 7 to 8 days. At each observation point, each mouse was scored for overall disposition, severity of abdominal breathing, presence of open-mouth breathing, activity level, presence of lethargy, and mortality. Mice which exhibited open-mouth breathing or were moribund were euthanized.
BoNT/A was iodinated using 125I-Bolton-Hunter reagent, purified, and assessed as previously described (19). Botulinum toxin type A was iodinated twice during the study, with an average specific activity of 1.76 × 102 Ci/mmol. Each batch of toxin was used over a period of ca. 60 days. Aliquots of material that were used in individual experiments over this period were subjected to additional chromatography (G-25) prior to use to ensure removal of free isotope and small polypeptide fragments.
Iodinated BoNT/A was administered intravenously to mice via the tail vein, and at various times thereafter specimens were obtained by retro-orbital bleeding (3, 19). Experiments were done with control animals (toxin only) and experimental animals (toxin preincubated with potential antagonists). In addition to monitoring the levels of toxin in blood, one set of experiments determined the extent of local toxin accumulation in tissues. Mice were anesthetized by administration of pentobarbital sodium (Nembutal) (50-mg/kg dose), after which the thorax was opened and the heart was exposed. A butterfly needle (23-gauge) was inserted into the left ventricle, an incision was made in the right ventricle, and the body of the animal was perfused with a heparinized solution of phosphate-buffered saline (45 to 60 ml). The liver and spleen were removed from the body, and the accumulation of iodinated toxin was determined.
A panel of BoNT/A-binding recombinant scFvs was identified that were derived from the B-cell repertoire of a sheep which had been potently immunized against BoNT/A, including boosts with increasing doses of active holotoxin. Sheep were used as the genetic source of scFvs, in part because they reportedly produce higher average affinity antibodies than do rodents (10). Several sheep were immunized by various methods and were tested for anti-BoNT/A titer and neutralizing activity. One sheep achieved a uniquely high antitoxin serum antibody titer in terms of both ELISA recognition of BoNT/A holotoxin (1:204,800) and neutralization (1 μl protected a mouse from the a dose equivalent to 10,000 LD50 of BoNT/A) and was chosen as the source of genetic material for the scFv library. The antibody repertoire was displayed as scFvs on phage and panned using a strategy to select for scFvs with high affinity for BoNT/A holotoxin.
The five unique anti-BoNT/A scFvs (scFvs 2, 3, 7, 8, and 21) producing the strongest signals by ELISA were expressed in bacteria, fused at the carboxyl end to a generic peptide epitope tag (E tag; GE Healthcare). The soluble scFv proteins retained the ability to recognize BoNT/A holotoxin and were each found to have a Kd of <20 nM by surface plasmon resonance (Table (Table1).1). The anti-BoNT/A scFvs were also tested for toxin-neutralizing activity by measuring their ability to inhibit BoNT/A intoxication of primary neuronal cells. Only scFv 2 was capable of producing detectable toxin neutralization at 35 nM (Fig. (Fig.1A),1A), and this scFv displayed a 50% inhibitory concentration (IC50) close to its 17 nM Kd for holotoxin (Fig. (Fig.1B).1B). scFvs 3 and 21 recognized isolated BoNT/A light chain, while scFv 7 recognized isolated native BoNT/A heavy chain (Table (Table1).1). Neutralizing scFv 2 did not recognize either of the isolated BoNT/A chains alone and lost binding for holotoxin following regeneration of the Biacore chip, suggesting that it may recognize a conformationally sensitive epitope.
Infusion of scFv 2 or 7 (20 μg) caused a short delay in the time to death in mice given a dose of BoNT/A equivalent to 10 LD50 (9 pg/g) (Fig. (Fig.2A).2A). While all control mice died within a day, 9 of 10 mice treated with either one of the scFvs survived longer than a day. When the scFvs were coinfused with an IgG1 anti-E-tag MAb (GE Healthcare) recognizing the epitope tag genetically fused to each scFv, the time to death was extended by several days for scFv 7, and with the toxin-neutralizing scFv 2, all five mice survived (Fig. (Fig.2A).2A). It was not necessary that the scFv and the MAb be administered at the same site. These results indicate that the antitoxin efficacy of the scFvs is improved by the presence of a bound antibody.
One possible reason that Ab bound to the antitoxin scFvs enhanced their efficacy is that the presence of the Fc domain improves Fc-mediated clearance of the toxin. Since Fc-mediated clearance involves a low-affinity Fc receptor, we hypothesized that antitoxin efficacy would be increased substantially by administering additional epitope-tagged scFvs, since this would increase the valency of Fc bound to the toxin. Vastly better Fc receptor recognition of “polymeric” Ab versus monomeric Ab is considered to be the mechanism by which immune complexes are efficiently cleared from serum in vivo without also rapidly clearing unbound antibody or simple antibody-antigen interactions (6). Soluble immune complexes have been shown to be cleared from serum within minutes, primarily through the liver by uptake into Kupffer cells (8, 13, 15).
To test whether in vivo protection against BoNT intoxication was improved with the administration of an increasing number of different tagged anti-BoNT scFvs together with an anti-tag MAb, a series of studies were performed. Figure Figure2B2B shows that mice given anti-E-tag MAb (5 μg), together with E-tagged scFv 2 or 7 (6 μg), rapidly died if exposed to a toxin dose equivalent to 100 LD50. When a pool containing 3 μg each of two scFvs (2 and 7) was administered, mice survived a dose of 10 LD50 without symptoms and had an extended time to death with a dose of 100 LD50 (Fig. 2A and B). If the pool contained 2 μg each of three scFvs (2, 3, and 7), together with anti-E-tag MAb, mice survived a dose of 100 LD50 of BoNT/A without symptoms, and four of five mice survived a dose of 1,000 LD50 (Fig. 2B and C) but did not survive a dose of 10,000 LD50 (not shown). This pool permitted only a small delay in time to death at 10 LD50 when it was not administered with the anti-E-tag MAb (not shown). A pool containing 2 μg each of three irrelevant scFvs administered with anti-E-tag MAb did not delay the time to death even at 10 LD50 (not shown). Finally, when a pool of 1.5 μg each of four scFvs (2, 3, 7, and 21) was administered with anti-E-tag MAb, all mice given a dose of 1,000 LD50 of BoNT/A survived (Fig. (Fig.2C)2C) and had no apparent symptoms of intoxication. This pool and dose were tested in a separate experiment (not shown), and all five mice survived a dose of 10,000 LD50, although they displayed moderate symptoms of botulism. Thus, in these experiments, mice in every treatment group received 6 μg total of tagged scFvs, and the antitoxin efficacy synergistically improved when the scFv pool contained increasing numbers of different scFvs together with the anti-E-tag MAb. Much higher doses of single scFvs with anti-E-tag MAb were protective only against challenges with 10 LD50 (Fig. (Fig.2A).2A). Increased numbers of different anti-BoNT/A scFvs, each binding at different sites on the toxin and bound by an anti-E-tag MAb, increase the valency of Fc domains bound to the toxin. The increased valency of Fc domains bound to BoNT/A is associated with a dramatic improvement in the antitoxin efficacy.
If the efficacy of the tagged binding agent antitoxin is related to the valency of anti-tag Abs bound to the toxin through the scFvs, then it should be possible to improve the potency by installing an additional epitope tag on one or more of the scFvs. To test this concept, a second copy of the E tag was genetically fused to the amino terminus of scFv 7 (scFv 7-E2). The presence of an additional epitope tag did not affect the binding of scFv 7-E2 to BoNT/A holotoxin (Table (Table1).1). The efficacies of different antitoxin scFv pools were investigated, with the only variable being the use of scFv 7 joined to either one or two copies of the E tag. As shown in Fig. Fig.3A,3A, treatment of mice with a pool containing 1 μg each of scFvs 3 and 7-E2, with anti-E-tag MAb, delayed the time to death by several days in animals receiving a dose of 100 LD50 of BoNT/A. An equivalent pool in which scFv 7 had only a single copy of the E tag delayed death less than half a day at this toxin dose. Using a pool of 0.67 μg each of three scFvs (2, 3, and 7 or 7-E2) also revealed substantial improvement in protection when scFv 7 contained two epitope tags. Both pools of three anti-BoNT/A scFvs with anti-E-tag MAb protected mice exposed to a dose of 100 LD50 without development of symptoms (Fig. (Fig.3A),3A), but only the pool containing the double-tagged copy of scFv 7 protected mice exposed to a dose of 1,000 LD50 from dying of intoxication (Fig. (Fig.3B).3B). In fact, none of the mice receiving this scFv pool developed visible symptoms with a dose of 1,000 LD50, and one of the five mice given this pool survived a dose of 10,000 LD50 (Fig. (Fig.3C).3C). Finally, mice given a pool of 0.5 μg each of four scFvs (2, 3, 21, and 7 or 7-E2) all survived a dose of 1,000 LD50 of BoNT/A without symptoms of intoxication, but only mice receiving the pool containing the double-tagged scFv 7 survived a dose of 10,000 LD50 (Fig. 3B and C). Clearly, the presence of an anti-BoNT/A scFv having a second copy of the epitope tag significantly improved the antitoxin efficacy of the scFv pools, giving approximately the improvement achieved by including an additional single-tagged anti-BoNT/A scFv.
A pharmacokinetic evaluation was performed to test directly whether the pool of tagged BoNT/A binding agents induced rapid serum clearance of BoNT/A in the presence of anti-tag MAb. In one experiment (Fig. (Fig.4A),4A), mice (n = 6 per data point) were injected with 125I-BoNT/A (5 ng/mouse; i.v.). At various times thereafter, animals were sacrificed and the levels of toxin in blood were determined. When the toxin was preincubated with a mixture of four anti-BoNT/A scFvs (2, 3, 7-E2, and 21) and anti-E-tag MAb prior to administration to mice, there was rapid clearance of toxin from blood compared to that of the control group. Within less than 10 min, the circulating titer of toxin had fallen almost 1 order of magnitude.
Antibody-mediated clearance is typically associated with accumulation of antibody-antigen complexes in the liver and/or spleen. To test this, groups of mice were administered toxin with or without preincubation with the anti-BoNT/A scFv mixture and anti-tag MAb for 30 min. The mice were then sacrificed and perfused, and their livers and spleens were excised and used to measure the amount of radioisotope. As shown in Fig. Fig.4B,4B, there was greatly enhanced accumulation of antibody-antigen complexes in livers in the treated versus control animals.
To better characterize the clearance-driven phenomenon, two independent experiments were performed in which groups of five mice were administered toxin (with or without conventional antitoxin serum) (18), an anti-BoNT/A scFv pool alone, anti-tag MAb alone, or a combination of the scFv pool and the MAb. The results from one experiment are shown in Fig. Fig.4C,4C, although both experiments produced equivalent results. The two major points to emerge from these studies were that (i) neither the scFvs alone nor the anti-E-tag MAb alone had significant ability to promote clearance—only the mixture produced a significant reduction in blood levels of toxin and a concomitant increase in liver accumulation of toxin—and (ii) the mixture of a neutralizing dose of scFvs and anti-E-tag MAb was comparable in efficacy to a neutralizing dose of polyclonal antitoxin serum.
To mimic a therapeutic situation, a pool of four anti-BoNT/A scFvs, each expressed with a single E tag, and anti-E-tag MAb were administered intravenously to mice either 2 or 4 h after they had received an intraperitoneal dose of BoNT/A equivalent to 10 LD50. The amounts of each scFv (1.5 μg) and of anti-E-tag (5 μg) administered to each mouse were selected as the doses maximally able to protect mice from a dose of toxin equivalent to 1,000 LD50 (i.e., higher doses of toxin were lethal) when coadministered with toxin as described above. As shown in Fig. Fig.5,5, treatment with this scFv pool completely protected mice from death when the pool was administered at 2 h postintoxication and delayed death when the pool was administered at 4 h postintoxication. Other groups of mice were given an intravenous dose of CDC anti-BoNT/A polyclonal antitoxin 2 or 4 h after intoxication with a dose of 10 LD50 of BoNT/A. Like the case for the scFv pool, the dose used had been found previously to be the smallest dose required to protect mice from a dose of 1,000 LD50 when coadministered with toxin. The polyclonal antitoxin did not protect mice that were given toxin at 4 h postintoxication, although when it was administered at 2 h postintoxication, this treatment protected 1 of 5 mice (Fig. (Fig.55).
Circulating polyclonal antibodies that recognize a disease-causing pathogen can elicit a therapeutic benefit through a variety of different effector mechanisms. One important mechanism is to attach multiple antibody Fc domains to a pathogen, and these are then bound by low-affinity Fc receptors in the liver and cleared from the circulation. Accelerated clearance in this way has the potential to improve disease when the target of the antibodies is a harmful biomolecule, such as a neurotoxin, that does not itself act on the liver cells as it is being removed from circulation and the body.
We show here that rapid serum clearance of the neurotoxin BoNT/A will occur if the toxin is bound by several different small binding agents that each contain a common epitope tag and if a single anti-tag monoclonal antibody is coadministered. A model for the proposed tagged-binding-agent clearance mechanism is shown in Fig. Fig.6.6. Several antitoxin binding agents become associated with the toxin at different sites. Since each binding agent contains one or two copies of an epitope tag, coadministration of an anti-tag MAb leads to the toxin becoming decorated with antibodies. We suggest that the presence of multiple antibodies bound to the toxin through the binding agents triggers toxin clearance through the liver by the same Fc-dependent process. Most likely, the mechanism involves Fc receptor-mediated clearance similar to that responsible for clearance of multimeric antibodies or immune complexes (6, 8, 13, 15). Also possible are complement-mediated mechanisms of immune complex clearance (22).
Several lines of evidence indicate that the toxin is removed from serum by mechanisms dependent on the antibody Fc domain. First, clearance is dependent on the presence of both the tagged antitoxin binding agent and the anti-tag MAb. The use of tagged binding agents lacking specificity for the toxin has no antitoxin efficacy, even in the presence of anti-tag MAb. Second, a second epitope tag on one of the toxin binding agents improves the antitoxin efficacy of the pool about the same amount as does the addition of another, different single-tagged binding agent. This implies that the additional epitope tag on the toxin binding agent results in an additional MAb Fc domain becoming associated with the toxin, just as would occur if another single-tagged binding agent were present. Finally, we show that the accelerated clearance occurs mostly or entirely through the liver, which is strongly suggestive of Fc-mediated clearance.
The result showing that the use of a single binding agent with two epitope tags improves the efficacy of the binding agent pool implies that the therapeutic efficacy could be increased further by engineering each of the binding agents to contain two epitope tags. The need for fewer binding agents would further reduce development and production costs for products employing this therapeutic strategy. While it is apparent that engineering two or even more epitope tags onto each of the binding agents will provide improved clearance efficacy, these binding agents will themselves become bound to two or more MAbs, and this will likely lead to their rapid liver clearance even when they are not bound to their pathogenic targets and thus reduce or negate further therapeutic benefit.
The potential of a pool of small antitoxin binding agents for therapeutic applications was further demonstrated by showing that treatments were successful even when the agent was delivered 2 and 4 h after intoxication with a dose of 10 LD50 of BoNT/A. Current polyclonal antitoxins are used primarily after it is shown that a patient may have been exposed to toxin, but before major symptoms have developed. This test showed that the efficacy of these agents was not dependent on premixing toxin and antitoxin (which is the standard laboratory method). For the postintoxication testing, we compared the use of small binding agent antitoxins to the use of polyclonal anti-BoNT/A antisera. We used doses of both treatments that were previously shown to maximally protect mice from a dose of about 1,000 LD50 when premixed with toxin and then administered (higher doses of toxin were lethal). Although the tagged binding agents appeared to be more effective than the polyclonal sera administered postintoxication, this is an imperfect comparison and cannot yet be interpreted to suggest that the new therapy is superior in terms of efficacy. Nevertheless, the study does indicate that tagged binding agent antitoxins may be at least equivalent to conventional antitoxins when used in a postintoxication therapy.
Agents that promote rapid clearance of toxins or other pathogenic molecules would provide therapeutic benefit in many disease situations beyond botulinum toxin exposure (e.g., exposure to other toxins and venoms, cytokines, etc.). The results presented here suggest that pools of tagged agents recognizing different epitopes on a pathogenic molecule, together with anti-tag antibodies, have promise as therapeutic agents to promote clearance of the pathogenic molecule from patients. Such binding agents should be much more rapid to identify and more convenient and economical to commercialize than polyclonal antisera or a pool of different MAbs (17). A wide variety of high-throughput strategies have been developed to rapidly identify small binding agents specific to defined targets (reviewed in references 11 and 12). Many of these binding agents are non-antibody-based and consist of scaffolds with excellent commercial and therapeutic properties, such as low production costs, long shelf lives, serum stability, low toxicities, and low levels of immunogenicity. Although the therapeutic strategy suggested here currently requires coadministration with an anti-tag MAb, the same MAb could be used in all therapeutic applications, and therefore it could be selected for optimal commercial and therapeutic properties, produced, and stockpiled. Furthermore, the MAb isotype (IgG1 in this study) could be altered if different isotypes provided improved therapeutic benefits. Alternatively, the antibody might be engineered for improved Fc binding to the Fc receptor or replaced altogether by a bispecific binding agent recognizing both the epitope tag and Fc receptors. In some situations, prior immunization with an epitope tagged immunogen could produce preexisting anti-tag polyclonal antibodies to obviate an anti-tag MAb. As such, we believe that this general strategy may find therapeutic applications in both biodefense and general medicine.
We are grateful to Jong-Beak Park and Jong-O Lee for performing cell-based assays for BoNT/A neutralization. We thank Michelle Debatis and Jacque Tremblay for excellent technical assistance and Kwasi Ofori, Karen Baldwin, and Chase McCann for their assistance with the animal studies. We also appreciate the assistance of Fetweh Al-Saleem and Denise Ancharski for their contributions to the pharmacokinetic experiments. We acknowledge Andrew Zhou and Kim Janda for their assistance with Biacore affinity analysis of some of the scFvs.
This project was supported in part with federal funds from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, Department of Health and Human Services, under contract number N01-AI-30050 and award number U54 AI057159 (S.T., C.B.S.), with additional support provided under DTRA contract 1-07-0032 (L.L.S.).
The content of this report is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Allergy and Infectious Diseases or the National Institutes of Health.
Editor: S. R. Blanke
Published ahead of print on 16 November 2009.