The development of new vaccines against Bacillus anthracis


Les Baillie Pathology, Biomedical Sciences, DERA Porton Down, Salisbury, SP4 0JQ, UK (e-mail:

1. Introduction, 609

2. The organism, 609

3. Disease in animals, 609

4. Disease in man, 610

5. Pathogenesis, 610

6. Virulence factors, 610

7. Preventative measures, 610

7.1. Antimicrobials, 610

7.2.Vaccines, 611

7.2.1. Animal vaccines, 611

7.2.2. Human vaccines, 611

7.2.3. UK licensed human anthrax vaccine, 611

7.2.4. New vaccines, 611

7.2.5. Third generation anthrax vaccines, 612

8. Into the post-genomic future, 612

9. References, 612


The disease anthrax is caused by a bacterium, Bacillus anthracis. Although primarily a disease of animals it can also infect man, sometimes with fatal consequences. Recent interest in the organism has centred on its ability to be employed as a biological weapon; the organism forms heat resistant spores which are easy to produce using commercially available technology and can infect humans via the aerosol route. At the time of the Gulf War it was reported that Iraq had produced and weaponized large quantities of anthrax spores (Zilinska 1997; As a consequence of this raised profile a considerable amount of effort has been focused on the development of new therapies that will meet the standards of the 21st century. Vaccination is the most cost-effective form of mass protection. While the first anthrax animal vaccine was developed by Pasteur in 1881, human vaccines did not emerge until the middle of the 20th century. The vaccines currently available provide effective protection but suffer from problems of standardization, are relatively expensive to produce, require repeated dosing and have been associated with transient side-effects. This paper will give a brief overview of the organism, its pathogenicity and the efforts being made to develop new vaccines.


The Bacillus genus is comprised of the Gram-positive aerobic or facultatively anaerobic spore-forming, rod-shaped bacteria. It is frequently convenient to class B. anthracis informally within the ‘B. cereus group’ which comprises B. cereus, B. anthracis, B. thuringiensis and B. mycoides on the basis of phenotypic reactions (Turnbull 1999).

However, genetical techniques have been developed which provide clear evidence that B. anthracis can be distinguished from other members of the genus (Keim et al. 1997, 2000). In practical terms the demonstration of virulence constitutes the principle point of difference between typical strains of B. anthracis and those of other anthrax-like organisms (Little and Ivins 1999; Turnbull 1999).


Anthrax is primarily a disease of herbivores, particularly the human food animals, and has a world-wide distribution. Prior to the advent of an effective vaccine it was a cause of heavy losses in cattle, sheep, goats, horses and donkeys in many parts of the world. In 1923 in South Africa it was estimated that between 30 and 60 000 animals died of anthrax (Sterne 1967).

In herbivores the disease usually runs a hyperacute course and signs of illness can be absent until shortly before death. At death the blood of an animal generally contains>  1 × 108 bacilli ml–1. Bacillus anthracis is regarded as an obligate pathogen; its continued existence in the ecosystem appears to depend on a multiplication phase within an animal host. Spores reach the environment either from infected animals and their products or as a consequence of the actions of man. In the wild it is thought that healthy animals acquire the disease by grazing on land contaminated with spores from infected animals (Turnbull 1996).

Over the past half century as a result of the combined benefits of vaccines, improved hygienic practices and the availability of antibiotics, anthrax has become rare amongst the livestock of most western countries. Unfortunately it still remains a major cause of livestock mortality in certain endemic countries which lack any efficient vaccination and control policy (Baillie 2000).


Man generally acquires the disease directly from contact with infected livestock (nonindustrial anthrax) or indirectly in industrial occupations concerned with processing animal products (industrial anthrax). While the incidence of human anthrax in Britain is low, significant outbreaks were reported from a number of countries mainly in Africa and Asia (Donganay 1990; Fujikura 1990). The outbreak in Zimbabwe which started in 1978 as a cattle outbreak resulted in over 10 000 human cases (Pugh and Davies 1990).

Three forms of the disease are recognized in humans: cutaneous, pulmonary (inhalation) and gastrointestinal. The pulmonary and gastrointestinal forms are regarded as being most frequently fatal due to the fact that they can go unrecognized until it is too late to instigate effective treatment (Dixon et al. 1999).


Despite the early understanding of the cause of anthrax, the pathogenic process in humans is understandably ill-defined. Animal studies have shown that the interaction of the organism with the macrophage is the key event in the disease process (Hanna 1997; Guidi-Rontani et al. 1999; Hanna and Ireland 1999). The spore form appears to require the environment of the macrophage to stimulate germination. The resulting vegetative form is able to survive the harsh environment of the macrophage and escape the cell by means as yet undetermined. It is known that the organism expresses a toxin called lethal toxin to which the macrophage is particularly sensitive. At low level this toxin stimulates the production of the shock-inducing mediators TNFα and IL-1β. At higher levels the toxin causes the death of the macrophage and the release of the cytokines into the blood stream. Death is probably due to a combination of massive bacteraemia and cytokine induced shock (Hanna 1997; Hanna and Ireland 1999).


It is likely that the organism expresses a range of virulence factors. The two major factors, a tripartite toxin and an antiphagocytic capsule, are encoded by genes carried on two plasmids, pXO1 and pXO2, the loss of either results in a marked reduction in the virulence (Little and Ivins 1999). The tripartite anthrax toxin is considered to be the major virulence factor. The three proteins that comprise the exotoxin are lethal factor (LF), oedema factor (EF) and protective antigen (PA). The toxins follow the A–B model: the A moiety comprises the catalytic subunits LF and EF, while the B moiety (PA) serves to translocate either EF or LF into the cytosol. Lethal toxin, the combination of LF and PA, is the central effector of shock and death from anthrax. The toxin contains a thermolysin-like active site and zinc-binding consensus motif HExxH and has been shown to act as a Zn2+ metalloprotease on a variety of substrates including peptide hormones and MAPKK (Duesbery et al. 1998; Pellizzari et al. 1999). It is now believed that the intracellular hydrolysis of important host protein substrates is responsible for the cellular toxicity associated with lethal toxin (Hanna and Ireland 1999). Oedema factor causes fluid loss through elevation of cellular cyclic AMP concentrations in the affected tissues. The contribution of EF to the infective process is ill-defined – it appears to inhibit the chemotactic response of polymorphonuclear leucocyte and may inhibit subsequent phagocytosis (Hanna and Ireland 1999; Keim et al. 2000). The B moiety is so named due to its role as the key protective immunogen in the current human vaccine. It binds to an as yet unknown cell surface receptor whereupon it is cleaved by the cell-surface protease furin to expose the A moiety binding site. Following proteolytic activation, PA forms a membrane-inserting heptamer that translates the toxic enzymes, lethal factor and oedema factor, into the cytosol (Petosa et al. 1997).


7.1. Antimicrobials

Prior to the advent of antibiotics a number of chemotherapeutic agents had been used to treat anthrax. In the early 1930s small doses of anti-anthrax horse serum were given intravenously and around the malignant pustule and in some cases massive doses of 1000 ml were administered as a single intravenous injection. In addition repeated large doses of neoarsphenamine were used, as were sulphonamides (Gold 1967).

With the advent of antibiotics the agent of choice for the treatment of anthrax became penicillin. Most strains of B. anthracis are sensitive to penicillin, erythromycin, chloramphenicol, gentamicin, ciprofloxacin and tetracycline. For patients allergic to penicillin any of these agents offers an effective alternative (Lightfoot et al. 1990).

It is important that chemotherapy is administered as early as possible, regardless of the antibiotic chosen. In cases of cutaneous anthrax, response to treatment is excellent provided it is started promptly. However, eschar formation may still occur due to the presence of toxin in the primary lesion.

Treatment of pulmonary anthrax by chemotherapy is usually ineffective, because the disease is rarely recognized before bacteraemia has developed and mortality is extremely high. The first stage of infection is often mistaken for influenza or bronchitis, and the second stage resembles cardiac failure or cerebrovascular accident.

If treatment is instituted early, pulmonary anthrax should respond to combined penicillin–streptomycin therapy (Gold 1967). Once treatment has commenced it is important to ensure that it is continued for prolonged periods. Animal studies have shown that even after 30 d of continuous treatment, infection can still occur if antibiotic cover is stopped. This is due to the ability of the spore to persist in the lungs for a considerable period of time. To counter this problem infected individuals should be vaccinated on commencement of antibiotic treatment so that treatment can be discontinued once protective immunity has developed (Friedlander et al. 1993).

7.2. Vaccines

Vaccination is the most cost-effective form of prophylactic treatment. For this reason a considerable amount of time and effort has been expended on developing safe effective animal and human vaccines.

7.2.1. Animal vaccines.

Studies on the vaccination of animals against anthrax date back to the end of the last century. Pasteur demonstrated protective immunization against anthrax in 1881 using a heat attenuated strain. Although effective the vaccine suffered from declining potency and troublesome variations in virulence which led occasionally to the death of animals. The search for a more effective and stable vaccine led to the development of the Sterne attenuated spore vaccine. This vaccine is based on an avirulent nonencapsulated strain 34F2 derived from the subculture of an isolate from a case of bovine anthrax. Since its introduction the vaccine has proved safe and extremely effective and has required little modification. While the vaccine is effective, repeated vaccinations are required for long-term protection; a single dose will only provide immunity for about a year (Turnbull 1991).

7.2.2. Human vaccines.

Immunization of humans with live spores has been limited to the former USSR and China (Knop and Abalakin 1986; Dong 1990). Western nations such as the UK and the US use nonliving subunit vaccines based on protective antigen due to concerns over the possibility of residual virulence (Turnbull 1991).

7.2.3. UK licensed human anthrax vaccine.

It had long been known that an extracellular protective antigen was produced when B. anthracis was grown in vivo (Belton and Strange 1954). The current UK licensed human vaccine (PL 1511/0058) consists of an alum-precipitated cell-free filtrate of culture supernatant from the Sterne noncapsulating strain of B. anthracis (Hambleton et al. 1984). The same strain is used to produce the live spore animal vaccine. Downstream processing consists of a filtration step which removes the bacterial cells along with some EF and LF. The sterile material containing PA, the essential protective immunogen, small amounts of EF, LF and other, as yet undefined bacterial and media derived proteins is alum-precipitated at pH 6·0. The presence of these additional proteins probably accounts for the transient reactogenicity seen in some individuals (Turnbull 2000).

7.2.4. New vaccines.

A considerable amount of research effort has been directed towards developing second generation human anthrax vaccines. Any new vaccine must be fully defined, produced from media free of animal derived products, capable of large-scale production, free of side- effects, and have undergone safety, efficacy, and clinical trials. In the medium term a subunit vaccine based on PA is the only candidate likely to receive licensing approval. As a consequence there have been numerous attempts to develop high level PA expression systems based on a variety of organisms; attenuated strains of B. anthracis (Belton and Strange 1954; Turnbull 1991; Ivins et al. 1998), B. subtilis (Ivins and Welkos 1986; Baillie et al. 1994, 1996, 1998), B. brevis (Oh et al. 1998), Vaccinia and Baculovirus (Iancono-Connors et al. 1990), Salmonella typhimurium (Coulson 1993) and E. coli (Vodkin and Leppla 1983; Singh et al. 1989). While none of these systems is ideal, to date, the best reported yields have been achieved with Bacillus spp.

The PA gene, which is plasmid encoded, is expressed well in B. subtilis achieving levels of expression higher than those obtained with the current B. anthracis based system (Ivins and Welkos 1986). A drawback of using bacilli is their ability to produce degradative proteases and initial attempts to produce fullsize rPA and develop downstream purification protocols were hampered by proteolytic degradation (Xu-Chu et al. 1991; Baillie et al. 1994,1996, 1998). This problem was overcome by transferring the PA gene into a protease deficient strain called B. subtilis WB600. With this host it has been possible to develop a research scale purification process which yields approx. 10 mg of pure rPA from a litre of culture (Miller et al. 1998). The ability of this recombinant protein to protect animals against challenge with anthrax has been demonstrated across a number of studies (McBride et al. 1998; Miller et al. 1998).

7.2.5. Third generation anthrax vaccines.

Ideally such a vaccine could be given orally, or intranasally, and would induce rapid immunity following a single dose. A number of technologies are currently under investigation for the development of third generation anthrax vaccines, two of which, live organism delivery systems and DNA vaccines, will be described.

Live vaccines are potential candidates for single dose, oral immunization. They consist of a microorganism into which the gene encoding the vaccine candidate has been introduced. Following delivery the organism either colonizes the mucosal surface or causes a limited infection during which the vaccine candidate is expressed and presented to the immune system: the Sterne animal vaccine represents such as vaccine. To date a number of live vector systems have been described for PA, attenuated B. anthracis (Barnard and Friedlander 1999; Cohen et al. 2000), B. subtilis (Ivins and Welkos 1986), Salmonella (Coulson et al. 1994), Lactobacillus (Zeger et al. 1999) and vaccinia (Bennett et al. 1999).

A disadvantage of live systems is the need to culture and store the organism prior to use. By contrast DNA vaccines represent an extremely simple and cost-effective technology.

They can broadly be defined as plasmid DNA expression vectors that when administered to an animal, result in expression of an antigen in situ leading to the induction of antigen specific immunity (Hassett and Whitton 1996). DNA based vaccines would offer a number of advantages over conventional subunit vaccines. DNA is relatively cheap to produce, unlike subunit vaccines which require extensive purification. They are heat-stable, thus removing the need for refrigeration, and are amenable to genetic manipulation, making it possible to design a vaccine that contains immunologically active sequences derived from multiple pathogens. DNA vaccines based on PA have been constructed and have been shown to be capable of inducing PA specific immune responses (Williamson et al. 1999) and of protecting mice against lethal toxin challenge (Gu et al. 1999).


An organisms biology is carried in its genetic material. Access to the chromosomal and megaplasmid nucleotide sequences of anthracis will free researchers from burdensome piecemeal acquisition of genetic information and allow them to study the biology of the organism with the whole-genome data set (Baillie and Read 2001). To date pXO1 (Okinaka et al. 1999) and pXO2 (GenBank Accession NC_002146, have been sequenced, and work to determine the chromosomal sequence of the Ames strain is underway currently at the Institute for Genomic Research. This sequence information can be searched through the TIGR website ( and will be updated regularly until the project is completed. With this information as a starting point, researchers will be able to identify new targets for vaccine and therapy development.