Antibodies for biodefense

Abstract

Potential bioweapons are biological agents (bacteria, viruses, and toxins) at risk of intentional dissemination. Biodefense, defined as development of therapeutics and vaccines against these agents, has seen an increase, particularly in the US following the 2001 anthrax attack. This review focuses on recombinant antibodies and polyclonal antibodies for biodefense that have been accepted for clinical use. These antibodies aim to protect against primary potential bioweapons, or category A agents as defined by the Centers for Disease Control and Prevention (Bacillus anthracis, Yersinia pestis, Francisella tularensis, botulinum neurotoxins, smallpox virus, and certain others causing viral hemorrhagic fevers) and certain category B agents. Potential for prophylactic use is presented, as well as frequent use of oligoclonal antibodies or synergistic effect with other molecules. Capacities and limitations of antibodies for use in biodefense are discussed, and are generally applicable to the field of infectious diseases.

Introduction

Biological warfare, understood as the intentional dissemination of biological agents such as viruses, bacteria or toxins to target numerous persons, can formally be traced to 1346, when plague-ridden human corpses were catapulted over the besieged walls of Caffa (now Feodosija, Ukraine).1 Much more recently, in WWII, plague was again utilized as a biological weapon (BW),2 and a biological bomb containing anthrax spores was developed, but not deployed.3 At present, BWs are banned by a United Nations' convention signed in 1972.4 However, with no verification protocol, this moral interdiction has not been sufficient to deter several countries from developing BWs.5 These same agents can also be weapons of choice for terrorist organizations6 due to their subjective, beyond their objective, impact. For these reasons the development of therapeutics and vaccines against BWs or biodefense, is an active field of research.

In 1901, Emil von Behring, a German military doctor, received the first Nobel prize for the initial development of polyclonal antibodies (pAbs) against infectious diseases such as diphtheria and tetanus. Sclavo's serum, another particular example of early pAb usage, increased the survival rate of cutaneous anthrax from 76–94%.7 At present, antibodies against infectious agents, but not against toxins, have been supplanted by antibiotics and antivirals. However, the problem of increasing bacterial resistance, with at least one known mechanism for each existing antibiotic class,8 requires development of new therapies at a time when only five new antibiotics have been approved by the Food and Drug Administration (FDA) between 2003 and 2007 and just two from 2008–2011.9 Given the early success of pAbs, recombinant antibodies, which are a successful class of therapeutic molecules,10–13 may constitute some of the new anti-infective molecules that are currently needed. Recombinant antibodies (rAbs) are defined in this review as antibodies which were selected or engineered, and expressed, utilizing DNA-based molecular biology techniques. Regarding the particular case of BWs, the risk of antibiotic resistance is acute because bacteria voluntarily disseminated could first be selected or engineered, for antimicrobial resistance. The development of antibodies may be a reasonable response to this risk; existing therapeutics, when they are effective, often act by other mechanisms and may act synergistically with antibodies. Antibodies may also be utilized for prophylaxis despite their relatively short half-life (∼three weeks for human antibodies), which can be increased using numerous methods.13 In effect, antibodies bring immediate protection that could be focused on personnel at risk of being exposed to BWs. The focused protection afforded by antibodies is advantageous compared with the protection elicited by vaccines, as vaccines must elicit an immune response to be effective and maintenance of this response may require booster injections. Further, rAbs may be administered in quantities that exceed that elicited by vaccines, and thus provide a higher level of protection, which is useful because BW exposure could involve elevated levels compared with natural exposure. In the particular case of botulinum neurotoxins, which are potential BWs, vaccination could be ethically disputable. RAb production is costly, but the number of doses needed is likely to be limited compared with the supply requirements for naturally encountered infectious diseases. This high production cost can be counterbalanced by the efficiency with which new rAbs might be isolated. It is thus no surprise that numerous efforts have been devoted to developing antibodies against BWs, particularly in the United States (US), following the 2001 anthrax letter attacks.

In this review, data corresponding to the best of our knowledge for rAbs at preclinical or clinical stages and intended for biodefense will be presented. When no such molecules were found, literature was searched for monoclonal antibodies (mAbs) at an earlier stage of development. To be complete, this study also presents pAbs approved for medical use. It is focused on antibodies developed against biological agents at the highest risk of being weaponized, known as bioterrorism categorized agents, as defined by the Centers for Disease Control and Prevention (CDC).14 Agents are classified under category A depending on their ease in transmission, high mortality and impact on social and public health infrastructures. Antibodies against category B agents will also be reviewed when these antibodies have proven efficient in vivo; the agents epsilon toxin of Clostridium perfringens, Chlamydia psittaci, Coxiella burnetii and Rickettsia prowazekii are excluded from discussion because no such antibodies exist. Each agent will be introduced and a potential role presented for antibodies, given the limits of other therapeutics and vaccines. Priority was given to results obtained on models utilizing the pulmonary route, by which many potential BWs are highly infectious. This route, against which physical protection is difficult, may threaten large populations.

Biodefense-related agents fall into the FDA “Animal Rule”15 where human efficacy studies are neither ethical nor feasible, and an appropriate animal model may clearly demonstrate efficacy. An animal model is appropriate if (1) it is supported by a “reasonably well-understood pathophysiological mechanism”, (2) if this mechanism is not well understood, the effect must be demonstrated in “more than one animal species” or “a sufficiently well-characterized animal model for predicting the response in humans”, (3) if “the animal study endpoint is clearly related to the desired benefit in humans”, and (4) if “the data…allows selection of an effective dose in humans”.16 This appropriate animal model is introduced in each section. Unless otherwise stated, models were utilized where all control animals died. The numbers (n) of animals tested are indicated in parenthesis. All indicated measurements, such as affinity, are presented as in the literature and may have been obtained with different conditions and equipment. With this review, we aim to present the state of the art and promote future development of antibodies for biodefense.

Category A Agents

Category A agents include Bacillus anthracis, Yersinia pestis, botulinum toxin, smallpox, Francisella tularensis and Ebola, Marburg, Lassa and Machupo viruses, and these agents cause the diseases listed in the present section.

Anthrax.

Bacillus anthracis is a gram-positive, spore-forming bacterium causing cutaneous, digestive or pulmonary anthrax. This latter form can cause a mortality rate as high as 100% when left untreated; however, during the more recent 2001 anthrax attacks, this rate was reported at 45% with treatment,17 due to a limited therapeutic window. B. anthracis produces several virulence factors that contribute to pathogenesis, in particular the lethal toxin (LT), which is composed of protective antigen (PA) and lethal factor (LF), the edema toxin (ET) composed of PA and edema factor (EF), and poly-γ-d-glutamic acid (γDPGA) capsule.18 The scientific consensus19,20 on the animal models for anthrax vaccines and therapeutics recommends rabbits and non-human primates (NHPs). The estimated lethal dose by an aerosol varies; however, the survival threshold in rabbits and NHPs is less than 10,000 spores.21 Post-exposure treatments involve antibiotics22,23 (fluoroquinolone, tetracycline or penicillin G), which must be continued for 60 days after inhalation,24 during which time the spores may still germinate. Analysis of the 2001 anthrax attack showed that out of 10,000 people potentially exposed to B. anthracis and treated, compliance to this long treatment regimen was only 40%.25,26

The US development of vaccines and therapeutics against anthrax includes BioThrax^® or Anthrax Vaccine Absorbed (AVA), by Emergent (Rockville, MD) which essentially consists of PA extracted from cultures of non-encapsulated B. anthracis. Given the existence of this approved vaccine, BW development in the US was mainly focused on therapeutic molecules, in particular to increase the therapeutic window and decrease treatment length. One method to reach both goals is administration of antibodies. Of note, to decrease the treatment length, the Advisory Committee on Immunization Practices has written that three doses of AVA at weeks 0, 2 and 4, administered with antimicrobial therapy, would be a beneficial post-exposure treatment due to the production of anti-PA antibodies.27,28 Four rAbs and one human pAb that target PA are currently undergoing clinical development or approval (Table 1).

Anthrivig™, or Anthrax Immune Globulin intravenous (AIGIV) by Emergent (Rockville, MD), is a pAb therapeutic developed from plasma of healthy donors vaccinated with Emergent's BioThrax^® Vaccine. Efficacy results are not published but tolerance was clinically verified under study designation NCT00845650 (www.clinicaltrials.gov).

Raxibacumab (Abthrax™) developed by Human Genome Services (Rockville, MD) is currently under final review for approval by the FDA.29,30 Raxibacumab is a human IgG1 isolated using phage display technology31 that binds to PA with an affinity of 2.78 nM.29 Raxibacumab prophylactically administered at 10 or 20 mg/kg subcutaneously (s.c.) 2 days prior, or concurrently at 40 mg/kg intravenously (i.v.), to rabbits (n = 12/group) challenged with 100 LD₅₀ aerosolized Ames spores provided 83, 83 and 100% survival, respectively.29 The 40 mg/kg dose prophylactically administered s.c. to NHPs (n = 10) two days prior to a challenge with 100 LD₅₀ aerosolized Ames spores was 90% protective.29 Treatment was also evaluated in rabbits and NHPs exposed to 200 LD₅₀ of aerosolized Ames spores. After PA was detected in the serum or following a 1.1°C rise in temperature, raxibacumab (40 mg/kg) was administered i.v. to rabbits or s.c. to NHPs, providing 44% (n = 18) and 64% (n = 14) survival, respectively.29 Raxibacumab's tolerance was also verified on human volunteers (n = 105).32 Human Genome Services was awarded a contract (HHS010020050006C) in 2006 to provide a total of 65,000 doses of Raxibacumab to the US Strategic Stockpile.

Valortim^®, also designated MAb-1303 or MDX-1303, isolated by Medarex (Princeton, NJ) and developed by PharmAthene, Inc., (Annapolis, MD), is in Phase 1 clinical studies (NCT01204866, NCT01265745). It is a human anti-PA IgG1 isolated from transgenic mice.33,34 Treatment was evaluated in rabbits and NHPs exposed to 200 LD₅₀ aerosolized Ames spores. In the rabbit model (n = 10), treatment with Valortim^® administered i.v. as two 1 mg/kg doses, given at 1 h and three days post-infection, provided 90% survival. Later treatment still provided 89% (n = 9) protection, but with an increased dose of 10 mg/kg given twice at 24 h and 120 h post-challenge.33,34 Treatment of NHPs (n = 6) demonstrated complete protection after a single intramuscular (i.m.) injection of 1 mg/kg Valortim^® given 1 hr post-challenge.33 Although a Phase 1 clinical study for Valortim^® has been completed for tolerance, combinational studies with antimicrobial treatments (NCT00964834 and NCT00964561) are apparently on hold or terminated due to a serious adverse event.

Anthim^®, previously designated ETI-204, is a chimeric IgG produced by Elusys Therapeutics, Inc. (Pine Brook, NJ) that binds to PA with an affinity of 0.33 nM.35 Its immunogenicity was decreased by DeImmunisation^®. A single 10 mg (∼4 mg/kg) dose of Anthim^®, prophylactically administered i.v. to rabbits (n = 8) 30–45 min prior to an exposure of 163 or 286 LD₅₀ aerosolized Ames spores, provided 100 and 88% protection, respectively. Treatment was evaluated in rabbits (n = 10) exposed to 172 LD₅₀ of aerosolized Ames spores, to which Anthim^® was administered as a single i.v. dose of 10 mg at 24 or 36 h post-infection and 80% or 50% survival rates were observed, respectively.35 Tolerance of Anthim^® both alone and in combination with ciprofloxacin have completed phase I clinical trials under respective study designations NCT00829582 and NCT00138411.

Thravixa^®, also designated AVP-21D9, is a human IgG1 developed by Emergent (Rockville, MD) from Epstein Barr Virus immortalized lymphocytes that were originally isolated from humans immunized with AVA. This antibody binds PA with an affinity of 82 pM.36 Protection and treatment by Thravixa^® was evaluated using both New-Zealand white and Dutch Belted Dwarf rabbits, yielding equivalent results. Ten mg/kg administered s.c. concurrently with an aerosol challenge of 87 or 100 LD₅₀ Ames spores were completely (n = 12/group) protective.37 Identical protection was observed against an intranasal (i.n.) challenge of 100 LD₅₀.38 Regarding treatment, s.c. administration of Thravixa^® (2 mg/kg) to rabbits at 0, 24 or 36 h following an aerosol challenge by 102 LD₅₀ Ames spores respectively provided 100%, 66% and 33% protection.37 Although not yet assessed for tolerance, in vivo synergistic effects were observed using combined treatments of Thravixa^® with a six day regimen of ciprofloxacin in a murine model. In particular, treatment of mice (n = 10) with 16.7 mg/kg Thravixa^® and 30 mg/kg/day ciprofloxacin resulted in complete survival against five nasally instilled LD₅₀ of Ames spores, versus 40% for the antibody alone and 60% for the antibiotic alone. Although complete survival against a similar challenge in a guinea pig model was not achieved, synergistic effects of Thravixa^® and ciprofloxacin were also demonstrated (80% versus 20% and 20% respectively for mAb and antibiotic alone).38

The current target of mAbs that are now in clinical trials against anthrax is exclusively PA, but a multi-targeted approach could be preferred. Although not in clinical trials, the latest generations of mAbs and antibody fragments directed against LF39,40 or γDPGA capsule41 promisingly neutralize in vitro or afford protection in animal models.

Plague.

Yersinia pestis is a gram-negative bacterium causing bubonic, septicemic and pneumonic plague. The latter form is of special biodefense interest as it is rapidly lethal.42 F1 is the dominant surface antigen of Y. pestis that relies, in particular, on the low-calcium response V (LcrV) antigen for virulence. The most appropriate animals for the study of plague are mice and African Green or cynomolgus macaque monkeys, as noted in the FDA workshop Animal Models and Correlates of Protection for Plague Vaccines (http://www.fda.gov/cber/minutes/workshop-min.htm). By the pulmonary route, the LD₅₀ of Y. pestis has been observed in different primate models as 100 to 20,000 aerosolized organisms.43,44 Although antibiotics45 (aminoglycosides, tetracyclines and chloramphenicol) are effective against pneumonic plague, they must be given within 24 h, and the separate appearance of two antibiotic-resistant strains exemplifies the need for additional therapeutics against Y. pestis.46 A recombinant F1+V protein (rF1V) vaccine is currently undergoing Phase 2 clinical trials (NCT01122784 and NCT00332956), but no antibody-based therapeutic has entered clinical trials.

Proof of antibody efficacy against Y. pestis was, in particular, provided by a murine anti-LcrV, mAb 7.3.47 Mice (n = 10) given mAb 7.3 (35 µg) by the intraperitoneal (i.p.) route, 4 h prior to an aerosolized challenge of 88 LD₅₀ (strain GB), demonstrated complete survival. When the same mAb dose was given 24 h post-infection, survival rates decreased to 80% (n = 10). Of note, aerosol delivery of 77.5 µg mAb 7.3 plus the same quantity of an anti-F1 mouse antibody, mAb F1-04-A-G1, 2 h post-infection increased survival to 82%, but against a lesser aerosolized challenge of 27 LD₅₀.48 Similar results with i.n. delivery were observed after the use of BA5, another murine mAb directed at LcrV.49,50 Challenges by the s.c. route gave results consistent with those observed after use of the pulmonary route.47,49

The rAbs m252 (anti-F1), and m253 plus m254 (anti-LcrV), are the only rAbs that provide protection in vivo in a mouse model of bubonic plague.51 These human IgGs were isolated from a naïve library and possessed sub-nM affinity. Mice (n = 6) challenged s.c. by 25–40 LD₅₀ (CO92 strain) were fully protected following the i.p. administration of m252 (500 µg/mouse) at 48 h post-infection. However, in this model, the survival rate dropped to 33% if m252 was administered earlier, at 24 h post-infection. A synergistic effect was seen when m252 was administered i.p. with m253 and m254 (500 µg of each) 24 h prior to a similar infection, with 83% survival compared to 33% for m252 alone. A combination of m253 and m254 failed to provide any protection.51

Future experiments studying the combined efficacy of antibiotics with antibodies are expected.

Botulism.

Botulism is a disease caused by the neurotoxins secreted by the gram-positive, spore-forming, anaerobic bacteria Clostridium botulinum.52 These bacteria produce seven serotypes of botulinum neurotoxins (BoNT), from BoNT/A to BoNT/G, whose sequences share a 34–64% identity,53 and sub-types have been described in reference 54. Four of these toxins (A, B, E and F) cause human botulism,55 but all seven can cause botulism in NHP inhalation models.56 BoNTs can be weaponized as aerosols as well as contaminate food and water supplies. In particular, BoNT/A is the most toxic substance on Earth, with a human LD₅₀ estimated at 1 ng/kg i.v., 10 ng/kg by aerosol and 1 µg/kg orally.52 A bivalent A/B H_C vaccine, rBV A/B produced by DynPort Vaccine Company LLC (Frederick, MD), has completed Phase 2 clinical trials (NCT00764634).57 The only two non-infant therapeutic products usable for biodefense are derived from equine pAbs; specifically, these are the heptavalent BoNT F(ab')₂ anti-toxin (HBAT) available from the CDC (Atlanta, GA),58 and the anti-ABE pAb Botulismus-Antitoxin Behring from Novartis (Basel, Switzerland). The neutralization capacities of anti-BoNT products are estimated in international units (IU), and one IU of anti-toxin protects a mouse against 10,000 mouse LD50 of A–F or 1,000 LD₅₀ of E BoNTs. For instance, 10 mL of HBAT represents 7,500 IU, 5,500 IU, 5,000 IU, 1,000 IU, 8,500 IU, 5,000 IU and 1,000 IU of anti-toxins A–G, respectively. Ten mL of Botulismus-Antitoxin Behring represents 7,500 IU, 5,000 IU and 500 IU of anti-toxin A, B and E, respectively. In NHP models, 143 IU/kg HBAT administered 48 h prior to 6 LD₅₀ of inhaled BoNT/A (assuming a NHP LD₅₀ by this route to be 300 mouse LD₅₀/kg) elicited full protection.59 Delaying treatment with the same dose of HBAT against the same challenge, at 46 h post-exposure, decreased survival of NHPs (n = 5) to 60%.59 The pAbs-derived products are well-suited to simultaneously protect against the various serotypes of BoNTs, but oligoclonal mixtures of recombinant antibodies have been developed to recapitulate that broad capacity. Oligoclonal combinations may also synergistically neutralize a single BoNT, as in the case of other toxins.60

Chimeric rAbs C25, S25 and the human rAb 3D12 bind to BoNT/A with 1.69 nM, 3.9 nM and 56.2 pM affinities, respectively.61 The potency of an equimolar combination of these antibodies was determined at 45 IU/mg. The synergistic neutralization effect of these antibodies against BoNT/A was demonstrated when a fixed quantity (50 µg in total, 25 µg of each paired antibody or 16.7 µg of each of three antibodies) was premixed with various quantities of BoNT/A and injected i.p. in mice (n = 10). Although paired mAbs afforded 90% protection against 1,000 LD₅₀, a mixture of the three antibodies was 100% protective against 7,500 LD₅₀, 80% protective against 10,000 LD₅₀ and approximately 50% protective against 20,000 LD₅₀.61

Several other studies have demonstrated protection with combinational antibody usage. Two IgG4s, C10 and 1B6, were generated from a synthetic human single-chain variable fragment (scFv) library. Each binds to the C-terminus of the heavy chain of BoNT/A (BoNT/A-Hc) with affinities of 64.6 pM and 23.8 nM, respectively. When combined, 25 µg of each mAb, pre-mixed with 20 LD₅₀ BoNT/A prior i.p. administration, completely protected mice (n = 6).62 Two IgGs against BoNT/A, 4LCA and 6A, were produced by hybridoma technology from humans immunized against pentavalent botulinum toxoid (PBT).63,64 4LCA and 6A bind to BoNT/A light and heavy chains, respectively, with affinities of 31 and 6.9 pM.63,64 Pre-mixing 50 µg of 4LCA and 6A each with 1,000 LD₅₀ BoNT/A prior to i.v. injection in mice (n = 6) resulted in complete survival.64 4LCA and 6A capacities have been further enhanced by attachment to a red blood cell-targeting protein.65

Recently, two human IgGs, 2B18.1 and 4E17.1, that cross-neutralize multiple serotypes of BoNT and are directed against the N-terminus of the heavy chains have been developed.66 They were isolated starting from humans immunized against PBT and with the use of shuffled libraries.67 The parental scFv of 2B18.1 binds to BoNT/A and B with respective affinities of 62.4 and 0.64 nM, respectively.66 Similarly, the parental scFv of 4E17.1 binds to BoNT/A, B, E and F with affinities of 0.09, 28, 0.23 and 16.8 nM, respectively. Twenty five micrograms of the rAbs 2B18.1 and 4E17.1 each, pre-mixed with 200 LD₅₀ of either BoNT/B or BoNT/E prior to i.p. injection in mice (n = 10), provided complete protection in both cases.

The National Institute of Allergy and Infectious Diseases (NIAID) has awarded a contract (HHSN272200800028C) to Xoma (Berkeley, CA) to produce mAbs against the major sub-types of BoNT A, B and E. Additionally, the European Union (EU) has established a collaborative AntiBotABE project to discover mAbs against the same toxins (www.antibotabe.com). The diversity of BoNTs represents a special challenge, and any single antibody able to neutralize multiple sero- or sub-types of BoNTs would significantly reduce the cost of any final oligoclonal product.

Smallpox.

Smallpox was declared eradicated by the World Health Organization (WHO) in 1980 after a successful global vaccination campaign, which was subsequently discontinued.68 The smallpox virus remains a potential BW precisely because a non-vaccinated population is vulnerable to this highly infectious agent, as a single inhaled virus particle can cause disease.69 Given this threat, live Variola virus is maintained for research purposes within only two secured labs, one in the US and the other in Russia. Smallpox virions may be encountered in two forms: intracellular mature virions (MV) or extracellular enveloped virions (EV).70 Strong antibody responses are observed in humans receiving the vaccine Dryvax^®,71 but its side effects have raised concern.72,73 A second generation vaccine, ACAM2000^®, was developed by Acambis Inc., (Cambridge, MA) and approved in 2007, although with restrictions that will not be discussed here in reference 74. Vaccinia Immune Globulin (VIG), a pAb extracted from vaccinated humans, is a product of limited potency,75,76 that is nonetheless approved for use against the potential side effects of the vaccine and reduces morbidity/mortality associated with the disease itself.75 Although still under debate, the WHO advisory committee on variola virus research recommends that approval of any prophylactic or therapeutic requires testing in NHP with fully virulent smallpox virus.77 Practically, however, given the restricted access to this virus, new antibodies are compared with VIG using a vaccinia strain. All antibodies described in this section were isolated from vaccinated or infected animals.

A humanized mAb, hB5RmAb, whose parental antibody was isolated from a rat, is directed against the B5 surface protein of EV.78 When mice were infected i.n. with 107 plaque forming units (PFU) of vaccinia, followed by treatment with the parental mAb (10 µg) administered i.p. at 5 h post-exposure, all survived and lost significantly less weight compared with controls.

Phage technology was utilized to isolate 8AH8AL, a chimeric chimpanzee/human rAb binding B5 with 0.6 nM affinity.79 Protection was compared to VIG utilizing the mouse pneumonia models, induced with 10⁵ PFU of vaccinia (western Reserve strain or VACV_WR) administered i.n. Decreasing i.p. quantities of 8AH8AL and VIG were administered to mice (n = 5) 24 h prior to the infection, and a dose of 22.5 µg 8AH8AL provided complete protection comparable to that elicited by 5 mg VIG. At 48 h post-infection, 90 µg 8AH8AL provided complete (n = 5) protection while 5 mg of VIG did not.79 In a second study utilizing the same approach, 6C was isolated and bound the EV protein A33 with 20 nM affinity (as measured with the parental Fab). In the same animal model, but 48 h after infection, rAb 6C was tested alone (90 µg) or in combination (45 µg each) with 8AH8AL. Full protection was observed in both cases, but weight loss was more limited with the combined treatment than with 6C only.80

Two human rAbs, hV26 and h101, were isolated from transgenic mice and bind to the H3 protein found at the surface of MV and to the B5 protein, respectively.81 Each mAb was administered i.v. and compared to VIG in severe combined immunodeficiecy (SCID) mice, utilizing New York City Board of Health Vaccinia virus (VACV_NYCBOH). Twenty five µg of each mAb administered to SCID mice one day prior to challenge against 5 × 10⁴ CFU VACV_NYCBOH elicited 50% protection, while all virus-treated controls given 1.25 mg VIG died.81

The WHO recommendation to test potential therapeutics against smallpox in NHP should be fulfilled for the best candidates.

Tularemia.

Francisella tularensis is a gram-negative intracellular bacterium of which several subspecies, indistinguishable by serological tests, are described. The subspecies tularensis or type A (which includes the SchuS4 strain, for instance) is the most virulent and most likely to be weaponized, as opposed to subspecies holartica or type B, from which the live vaccine strain (LVS) is derived.82 Tularemia can arise from as few as ten organisms administered either s.c.83 or by aerosol,84 with multiple clinical manifestations. Although there is no consensus on a small animal model, several studies85 suggest murine models utilizing the SchuS4 strain are capable of satisfying the FDA animal rule. F. tularensis is generally susceptible to antibiotics (e.g., tetracycline, fluoroquinolones, aminoglycosides), but susceptibility testing is recommended,86 and a new formulation of LVS used as a vaccine is currently in a clinical Phase 2 study (NCT01150695). In murine models, vaccination by LVS fully protects against an intradermal (i.d.), but not aerosolized, challenge of 1,000 LD₅₀ SchuS4,84,87–89 Prophylaxis and treatment by antibodies were first shown by transfer of immune serum against an aerosolized challenge by F. tularensis LVS,90 but only increased mean time to death (MTD) was observed after a challenge with SchuS4. All mAbs that protect against F. tularensis are, thus far, of murine origin.

MAb 12, an IgG2a, was developed after immunization by LVS and has been tested against both LVS and SchuS4 strains.91 Three doses of 50 µg each, prophylactically administered at days −1, 0 and +1, provided complete survival in mice (n = 4) against an intradermal (i.d.) challenge by 7 × 10⁷ CFU of LVS. When the same three doses were administered therapeutically at 1, 3 and 5-days after infection by the same challenge, survival decreased to 50%. Regarding type A, prophylactic administration of MAb 12 did not elicit protection, but doubled MTD against a 24 CFU challenge of virulent Schu24.91 A murine anti-LPS IgG2a, mAb 3, is the only mAb tested against an i.n. challenge of F. tularensis.92 Treatment was evaluated in mice challenged against 2 × 10⁴ CFU LVS, resulting in 100% survival following the i.n. administration of mAb 3 at 50 µg within minutes of challenge (n = 5), or following the i.p. administration of mAb 3 (200 µg) at 1 h post-challenge (n = 5).

Antibodies can completely protect against LVS but not against SchuS4. This might be due to different virulence mechanisms between the strains,93 explaining why effective protection against LVS (belonging to type B) does not guarantee protection against type A.

Viral hemorrhagic fevers (VHF).

Viral hemorrhagic fevers (VHF) are caused by four virus families: Filoviridae, Arenaviridae, Flaviviridae and Bunyaviridae. In particular, Category A agents include Ebola and Marburg of the Filoviridae family, as well as Lassa and Machupo of the Arenaviridae family. Many VHF agents are highly infectious by aerosol, and mortality rates may be greater than 90% during natural outbreaks.94 A consensus on the appropriate animal models of diseases caused by Filoviruses has not yet been reached due to symptom discrepancies between models and humans, but guinea pigs and NHPs serve to date as the primary models for these diseases.95 Ribavirin may be utilized to treat VHF, but only if caused by arenaviruses and bunyaviruses. Several DNA plasmid vaccines against Ebola and Marburg are in Phase 1 clinical studies (NCT00072605, NCT00374309, NCT00997607, NCT00605514). Immunotherapy against VHF was first demonstrated by employing crude blood transfusions during the 1995 Kikwit Ebola outbreak, with only one (n = 8) patient death following this treatment, as compared to 80% without.96 A working consensus regarding VHF states that passive immunization strategies using rAbs should be pursued in the future.97

Given the complexities associated with animal models for these agents, which require BSL-4 laboratory containment, only a few antibodies directed against VHF have been tested in vivo. The first of these mAbs was KZ52, a human rAb developed from survivors, binds to the glycoprotein (GP) of Ebola Zaire and effectively neutralizes the virus in a plaque reduction assay.98,99 Passive administration of KZ52 (25 mg/kg) in guinea pigs (n = 5) provides complete protection 1 h prior and 80% survival 1 h after s.c. challenge against 10,000 PFU of Ebola Zaire Mayinga.100 Two i.v. doses of KZ52 (50 mg/kg) given to macaques (n = 4), one day before and 4 days after an i.m. challenge with 1,000 PFU Ebola Zaire (Kikwit), failed to affect disease progression.101 A later study compared the in vitro inhibition mechanisms of KZ52 to J3PK11, developed from a rhesus macaque that survived an Ebola Zaire infection,102 and found them to be distinct.103 Although J3PK11 has only been tested in vitro, this result demonstrates that viruses responsible for Ebola and possibly other VHFs have multiple physiopathological mechanisms which may depend on the model utilized.

Category B Agents

Category B agents include Brucella, Burkholderia, ricin, Staphyloccal enterotoxin B alphaviruses, as well as food and water safety threats, and these agents cause the diseases listed in the present section.

Brucellosis.

Brucellosis is an anthropozoonosis caused by intracellular, gram-negative bacteria of the genus Brucella and B. melitensis is the major cause of the 500,000 annual cases of human brucellosis. These microbes are highly infectious, as aerosols of B. melitensis in NHP require only 10² CFU to cause disease.104 Mice are the primary animal model to characterize and test therapeutics against brucellosis.105 Antibiotics (tetracyclines, aminoglycosides, rifampicin and streptomycin) are generally effective, often combined in 6-week treatments, but the risk of relapse is as high as 30%.106 There is no licensed human vaccine for brucellosis. The last expert committee of the Food and Agriculture Organization of the United Nations/WHO in 1986 recommended development of antibody therapeutics towards brucellosis;107 however, the only therapeutic mAbs developed to date are of murine origin.108–110

Prophylactic i.p. administration of anti-LPS IgG3 mAb 6B3 (100 µg) to mice (n = 3), 24 h prior to i.p. challenge with 3.6 × 10⁵ B. melitensis 16M, provided complete protection. In a similar challenge but with 1.8 × 10⁵ CFU B. abortus, 50 µg of anti-LPS IgG3 mAb 2C8 fully protected the animals (n = 3).110

Melioidosis and glanders.

Glanders, a disease that primarily affects equids and melioidosis are difficult to discriminate clinically in humans. They are caused by Burkholderia mallei and Burkholderia pseudomallei, respectively.111,112 In vivo studies using mouse models have been considered the most effective due to their adaptability to investigate different aspects of disease,113 with BALB/c being the most susceptible to melioidosis.114 The LD₅₀ in murine models is strain-dependent, but has been reported as 1.6 × 10³ CFU for B. mallei and between 20–320 CFU for B. pseudomallei.115–117 Antibiotic resistance poses an issue to treat these diseases and carbapenems, as well as the trimethoprim-sulfamethoxazole association, may be recommended.118 Currently, no vaccine effective against either of these bacteria exists.112

The murine anti-LPS IgG2a mAb 1G2-1D3 was developed against Burkholderia mallei.119 The i.p. administration of 1G2-1D3 (1 mg) to mice (n = 6) 18 h prior to challenge with 20 LD₅₀ (1.9 × 10⁴ CFU) aerosolized B. mallei strain China 7 was completely protective. Under a similar challenge, the same administration of mAb 1G2-1D3 to mice but 18 h post-infection failed to affect disease progression.119

The murine IgG3 mAb Ps6F6 binds B. pseudomallei exopolysaccharide.116 Three i.p. injections (3.5 µg each) of mAb Ps6F6 to mice (n = 85/group), on day −6, −3 and −2 prior to i.p. challenge against 490 CFU (1.4 LD₅₀) B. pseudomallei strain 6068 VIR, failed to achieve greater than 40% survival when 10% of controls survived.116 In another study, seven murine mAbs demonstrated passive protection against the B. pseudomallei 4845 strain.120 Protection was evaluated using 40 µg of seven antibodies (6 µg of three anti-exopolysaccharides and of three other mAbs, plus 4 µg of an anti-LPS) as an oligoclonal cocktail administered i.v. to mice (n = 10). These antibodies, given 4 h before challenge, were completely protective against 10⁴ CFU (250 mean morbidity or MD₅₀) by the i.p. route, but did not protect against a higher dose of 10⁶ CFU.120

Intoxication by ricin.

Ricin can be easily extracted from Ricinus communis, a plant cultivated world-wide, also known as castor oil plant. This protein toxin consists of a B-subunit (RTB) that binds sugars on the cell surface for cytosolic entry of an A-subunit (RTA) that inhibits protein synthesis. It is lethal by the oral route, but 1,000-fold more potent by the pulmonary and parenteral routes. Ricin has an aerosolized LD₅₀ of 3–5 µg/kg in the mouse and 5.8–15 µg/kg in NHP models.121,122 Two potential vaccines currently in Phase 1studies (NCT01317667, NCT00812071), protect against ricin intoxication. For treatment against ricin, there are supportive, but no specific, measures. However, in a proof of concept study utilizing direct inhalation for intoxication and treatment, pAbs against ricin were completely protective 20 min post-challenge.123

Two chimeric IgGs whose parental mAbs were tested in a murine model have been developed. The murine mAb 4C13,124,125 chimerized as IgG c4C13,126 was administered at a dose of 100 µg/mouse (n = 4) by i.p. 30 min after i.p. challenge with 2 µg ricin (10 LD₅₀), and all animals survived.125 A second murine mAb, RAC18, was evaluated in an aspiration model where ricin and the antibody are instilled in the oropharynx.127 Mice (n = 10) challenged with 16 µg/kg (∼3 LD₅₀) ricin before the administration of 50 µg of antibody completely survived if the antibody was administered 4 h post-intoxication, and 60% and 50% survived if the antibody was administered at 18 and 24 h post-intoxication, respectively.127

The latest antibodies neutralizing ricin are more human-like, as with the macaque scFv anti-RTA 43RCA.128

Intoxication by Staphylococcal Enterotoxin B (SEB).

Staphylococcus aureus is a gram-positive bacterium that produces many different virulence factors. One of these is Staphylococcal enterotoxin B (SEB), a superantigenic toxin, and just one of more than twenty-five different staphylococcal enterotoxins (SEs) characterized to date. SEB, like many other superantigens, stimulates T cells to release levels of pro-inflammatory cytokines that can cause shock and death. Different animal models for investigating SE-induced shock, as well as potential therapeutics and vaccines, include multiple strains of mice, macaques, piglets, ferrets and shrews.129–141 Readouts of intoxication include fever, vomiting, diarrhea, as well as lethality.

Intravenous immunoglobulin (IvIg), collected from humans, can protect against SEs and related toxic shock syndrome toxin-1 (TSST-1).142,143 Humans are naturally in contact with S. aureus that grows on the skin, mucosal surfaces or in food items, and can seroconvert to S. aureus antigens that include the SEs. Experimentally, passive administration of pAbs (10 mg/kg), 4 h post-challenge, completely protected macaques (n = 4) against five aerosolized LD₅₀ of SEB.144

Two murine-human chimeric antibodies that bind distinct epitopes have recently been described but only tested in vitro.145 The mAb HuMAb-154 neutralizes SEB in vitro (proinflammatory cytokine release from peripheral blood mononuclear cells) and in vivo (mouse lethality).146 Screening of antibodies was done with a recombinantly-attenuated form of SEB used as a vaccine in various animal models.137,138,147,148 HuMAb-154 binds to SEB with an affinity of 0.29 nM and cross-reacts with SEA, SEC1 and SED by ELISA. Pre-mixing increasing concentrations of SEB with 500 µg of HuMAb-154 prior to i.p. administration to mice (n = 5) resulted in complete protection against 5 and 10 µg (25 and 50 LD₅₀, respectively) SEB, and 40% protection against doses as high as 100 µg (500 LD₅₀). A similar dose of HuMAb-154, administered to mice (n = 30) as a therapeutic at i 0, 0.5 and 1 h post-exposure to 10 µg (50 LD₅₀) SEB elicited 86, 50 and 13% survival, respectively.146

A human phage-displayed library was used in another study to generate bivalent Fabs and full length IgG.149 Pre-mixing 10 µg of the full length mAbs FL10 or FL9 with 2.5 µg SEB, prior to i.p administration to mice (n = 6) elicited 68 and 17% protection, respectively.149

In regards to antibody therapy against SEB in a biowarfare/bioterror scenario, none of mAbs discovered have been used in NHPs challenged with an SEB aerosol. Such experiments represent a high stringency test necessary to establish efficacy of therapeutic antibodies against SEB.

Encephalitis caused by alphaviruses.

Viruses of the Alphavirus genus (Togaviridae family) cause encephalitis, and of these, Venezuelan equine encephalitis virus (VEEV) has been more extensively studied and weaponized.150 The most relevant animal models for studying VEEV are considered mice and NHPs,151 with the LD₅₀ of VEEV in mice estimated as 1–30 PFUs. Two vaccines are currently in Phase 2 clinical trials, the attenuated TC-83 (NCT00582504) and the inactivated C-84 as a booster to TC-83 (NCT00582088).152,153 Supportive care is available for VEE, but there is no specific treatment.

The humanized mAb Hu1A3B-7 (IgG1 isotype) is broadly specific to VEEV subtypes, neutralizing type IA/B (Trinidad Donkey or TrD), type II (Fe37c) and type IIIA (Macambo BeAn8) in vitro.154 Twenty five micrograms of Hu1A3B-7 administered i.p. to mice (n = 10) 24 h prior to s.c. challenge with 100 LD₅₀ TrD strain, provided complete protection. Treatment was evaluated in mice (n = 10) exposed to 10 and 100 aerosolized LD₅₀ TrD: 100 µg Hu1A3B-7 administered 24 h post-challenge yielded 100% and 90% survival, respectively.154 Another humanized mAb, Hy4 IgG, binds to the E2 glycoprotein and is protective in murine models.155 Five hundred micrograms of mAb Hy4 administered i.p. at 24 h pre-challenge against 100 i.n. MD₅₀ TrD 1350 PFU) protected 80% of the animals (n = 10). After i.p. challenge by 100 LD₅₀ TrD, the i.p. administration of 10 µg mAb Hy4 at 1 and 24 h post-infection were 90% (n = 10) or 75% (n = 20) protective.155 A third mAb, hu1A4A1IgG1-2A, binds to the E2 glycoprotein with an affinity of 3.9 nM.156 Treatment of mice (n = 8) with hu1A4A1IgG1-2A was evaluated against the s.c. administration of 30–50 PFU TrD and completely protected when administered i.p. (50 µg) at 24 h post-infection, but the same dose failed to protect at 72 h post-challenge. A prophylactic administration of mAb at 24 h pre-infection resulted in complete protection.156

Recently several human antibody fragments have been isolated but have yet to be tested as full IgGs in vivo.157,158 Of note, the particular variability of such RNA viruses is of concern because it favors the emergence of resistance and consequently several therapeutic molecules, including Abs, may be required.

Diseases caused by food and water safety threats.

Food and water safety threats include a diverse class of organisms that present both biodefense as well as natural outbreak concerns, exemplified by the European E. coli outbreak in May 2011. Thus far, only shigatoxins and Vibrio cholerae have been targeted by antibodies evaluated in vivo.

The chimeric mAbs HαStx1 and HαStx2 (Shigamabs^®, also called cαStx1 and cαStx2), were derived from murine mAbs 13C4,159 and 11E10 160 and respectively target the Stx 1 and Stx 2 shigatoxins produced by E. coli.161–163 They are in Phase 2 clinical trials (NCT01252199) by Thallion Pharmaceuticals (Montréal, Québec) in collaboration with Laboratoires Français de Fractionnement et des Biotechnologies (LFB; Les Ulis, France).164 Protection was tested against Stx1 toxin with HαStx1 administered in two i.p. injections of 2.05 µg of each, at 24 h prior to challenge and simultaneously to challenge; mice (n = 5) were completely protected after i.v. intoxication with 20 LD₅₀ of crude Stx1.161 Using the same administration scheme but against Stx2 toxin, HαStx2 was administered in two 0.5 µg injections against 10¹⁰ CFU of E. coli strain 86–24 (producing Stx2), or two 116 µg injections against 10³ CFU of E. coli strain O91:H21 (producing Stx2d161), resulting in complete survival against each strain.161

The mAb 5C12, from a transgenic mouse, binds Shiga toxin Stx 2 A-subunit with an affinity of 0.85 nM.165 Treatment against this toxin was evaluated in piglets orally challenged with 10¹⁰ CFU of shigatoxigenic E. coli (STEC) 86-24 (producing Stx2), with 3 mg/kg 5C12 administered i.p. at 24 and 48 h post-challenge, resulting in 78% (n = 9) and 100% (n = 4) survival, respectively.165 Mice (n = 10) orally challenged with 10¹⁰ CFU of another shigatoxigenic E. coli, B2F1 (producing Stx2) were treated with 2.1 mg/kg mAb 5C12 administered i.p. at 0, 12, 24 and 48 h post-challenge, resulting in 80, 70, 90 and 60% protection, respectively.166

The murine anti-LPS mAb 72.1 is directed against Vibrio cholerae and protects against both the Ogawa (O395) and Inaba (569B) strains.167 Fifty µg of mAb 72.1 mixed with 10⁸ bacteria of each strain, given by oral gavage to infant mice (n = 5), provided complete protection in both cases.167

Discussion

Recombinant antibodies represent a rapidly emerging class of potent therapeutics, but, contrary to the historical use of pAbs, most rAbs currently marketed do not target infectious agents. Palivizumab (Synagis^®), used to prevent infections caused by respiratory syncytial virus, but with incomplete success, is the single exception.168 The limited development of antibodies against infectious agents may be explained by existing antibiotics or antivirals, but these molecules have their own limitations. For instance, antibiotic resistance is a particularly high risk regarding bioweapons, and the available therapeutic window is often narrow. Such limitations explain that, perhaps particularly after the deliberate dissemination of B. anthracis spores in the US in 2001, antibodies against anthrax and other bioweapons were increasingly developed. Regarding anthrax, the current stockpiling of an effective rAb,29 and existence of three competitors,33,35,37 showcase the capacity to isolate efficient rAbs when microbial pathogenesis is understood. However, the situation regarding anthrax has no equivalent so far, perhaps because oligoclonal rAbs may be necessary to effectively target a single agent, such as Y. pestis51 or B. pseudomallei,120 thus increasing the complexity of antibody development. This may also be the reason why, despite impressive scientific results, no combination of antibodies targeting the various botulinum toxins has yet entered clinical trials. In fact, the narrow specificity of antibodies limits their efficacy, although it conversely limiting their reactivity with human tissues, which confers an advantage regarding clinical tolerance. Antibodies developed against smallpox represent one of the several possibilities of antibody prophylaxis.78–81 Indeed, antibodies are often more efficient for prophylaxis versus therapy of infectious diseases, as discussed in this review and also exemplified by palivizumab.168 There are extreme cases, such as brucellosis, where only prophylactic use is effective.110 Of note, while early treatment is almost always better, fast diagnosis is imperative, and antibodies certainly have a role in biodefense for the development of rapid tests.42,169,170

Targeting toxins, viruses or intra-cellular bacteria before internalization is more effective for antibodies, which usually do not enter cells, but antibodies might offer possibilities that are otherwise scarcely exploited. These possibilities include administration by aerosol, exemplified by antibodies targeting ricin,123 which could serve as an example for new clinical developments. Additionally, the effective use of antibodies against Alphaviruses may exemplify antibody capacity to cross the blood-brain barrier,171 and could prompt clinical developments targeting other central nervous system diseases.

All results were not as positive as expected, such as in the case of protection against Ebola virus,103 which is probably an example of where more fundamental research is needed. In effect, antibody inhibition of infectious mechanisms may be model-dependent and perhaps not always relevant to humans. More research might also be needed regarding the most virulent type A strains of F. tularensis, against which no antibody effectively protects. Being efficacious against several extremely virulent pathogens, antibodies for biodefense highlight their capacities against infectious diseases,172 and thus should lead investigators toward new clinical developments and research. Antibodies do have intrinsic limits, but these should not be confused with the limits of our knowledge.

Abbreviations

BW	=	biological weapons
mAbs	=	monoclonal antibodies isolated by cellular techniques
pAbs	=	polyclonal antibodies
rAbs	=	recombinant antibodies
i.m	=	intramuscular
i.p	=	intraperitoneal
s.c.	=	subcutaneously
i.v.	=	intravenously
i.n.	=	intranasal
MD50	=	mean morbidity
MTD	=	mean time to death
FDA	=	Food and Drug Administration
CDC	=	Centers for Disease Control and Prevention
WHO	=	world health organization

Figures and Tables

Table 1 Clinical and developmental status of anthrax antibodies targeting biodefense agents

International non-proprietary name, product name or code	Targeted antigen	Antibody type (Isolation)	Clinical phase or approval year	US Patent reference number
Raxibacumab/ABthrax™,a	anti-PA	Recombinant IgG1 mAb (naïve library)	FDA review	601351
Valortim^®/MAb-1303b	anti-PA	Human mAb (transgenic mice)	Phase 1	7456264
Anthim^®/ETI-204c	anti-PA	Chimeric deimmunized mAb (murine origin)	Phase 1	7446182
Thravixa™/AVP-21D9d	anti-PA	Human mAb (hybridoma)	Phase 1	7438909, 7442373
Anthrivig™/Anthrax Immune Globulin (AIG)	anti-PA	Polyclonal antibody from AVA human plasma	Phase 1/2	N/A

a Subramanian et al. 2005 and US Patent 7601351.

b Vitale et al. 2006 and US Patent Number 7456264.

c Mohamed et al. 2005 and US Patent Number 7446182.

d Peterson et al. 2007 and US Patents 7438909 and 7442373.

Acknowledgments

J.W.F. II is part of a formal Engineer-Scientist Exchange Program established between the governments of France and the USA. All opinions are solely those of the authors and do not reflect official government views.