The causative organism, Bacillus anthracis, is a spore-forming, rod-shaped bacterium that inhabits the soil. Although the causal relation between the organism and anthrax has been known for more than a century, since the time of Koch and Pasteur, the specific factors responsible for the virulence of the organism have been identified and characterized during the last 30 years. The precise molecular basis for virulence has been elucidated only during the past decade.

Fully virulent strains of Bacillus anthracis possess two unique virulence factors: a poly-D-glutamic acid capsule that inhibits phagocytosis [1] and a tripartite toxin composed of protective antigen, edema factor, and lethal factor [2]. Capsules are produced by virulent strains of Bacillus anthracis growing in vivo and by cells grown on media containing serum or bicarbonate or both and incubated in a CO₂-enriched atmosphere.

The existence of an anthrax toxin was first demonstrated in 1955 in experiments that showed that injection of sterile plasma from infected guinea pigs resulted in local edema and death. Studies by American and British investigators during the ensuing decade showed that the toxin contained three separate components. The individual toxin components have no known biological effects when administered alone, but edema factor injected with protective antigen into the skin of rabbits or guinea pigs causes local edema, and protective antigen injected with lethal factor into rats causes death in as little as 60 minutes. Protective antigen, so called because its injection into experimental animals results in protective immunity, binds to cell-surface receptors to produce an uptake system that can be used by both the edema factor and lethal factor to gain access to the cytoplasm. "Edema toxin" (edema factor and protective antigen) and "lethal toxin" (lethal factor and protective antigen) thus resemble the A-B enzyme-binding structures characteristic of many well-studied bacterial toxins. After protective antigen, which is analogous to the B chain, binds to a specific membrane receptor on the surface of a eukaryotic cell, it is cleaved at a single site, exposing a binding site for the other toxin component. The membrane-bound fragment of the protective antigen then binds to the edema factor or lethal factor and mediates the entry of the active moiety into the cell.

The edema factor has been found to be a calmodulin-dependent adenylate cyclase that elevates cyclic adenosine monophosphate (AMP) levels approximately 200-fold greater than normal in Chinese hamster ovary cells. Local edema, a typical sign of anthrax, may be directly related to adenylate cyclase activity associated with the edema factor. The increase in intracellular cyclic AMP caused by this toxin may lead to edema in a manner analogous to the loss of water into the intestinal lumen caused by cholera toxin, which also increases intracellular cyclic AMP [3]. Edema factor and the protective antigen also inhibit phagocytosis of anthrax bacilli by polymorphonuclear leukocytes, and this effect may further increase host susceptibility to anthrax. The dependence of edema factor activity on calmodulin, a substance found only in eukaryotic cells, suggests that the edema factor did not evolve from a bacterial enzyme but from a eukaryotic adenylate cyclase, the gene for which was adventitiously transferred into Bacillus anthracis and retained because it made the bacteria more virulent [3].

The mechanism of action of lethal factor is poorly understood, but it is lethal for many species of experimental animals and is assumed to be the major factor causing death in anthrax. No enzymatic activity has yet been associated with the lethal factor, and the nature of its intracellular target is unknown.

Virulent strains of Bacillus anthracis contain two large plasmids, pX01 and pX02 [1, 4, 5]. Both plasmids are required for full pathogenicity, and strains that contain only one of these plasmids are avirulent. The plasmid pX01 encodes all three components of the anthrax toxin, and pX02 encodes the poly-D-glutamic acid capsule. These facts provide the basis for effective vaccines, the first of which was developed by Pasteur in 1881 and used in a brilliantly successful field trial in sheep at the village of Pouilly-le-Fort. The avirulent Sterne vaccine strain, which is pX01⁺/pX02^-, produces toxin but no capsule and is used effectively as a live veterinary vaccine [6]. The nonencapsulated strain used in production of the acellular vaccine licensed for human use in the United States produces primarily protective antigen [6]. The heat-attenuated Pasteur vaccine strains form capsules but cannot produce toxin. Pasteur probably cured his strains of plasmid pX01 by heat attenuation to produce his vaccine for immunization of cows and sheep. Although the Pasteur-type vaccine provides a much lower level of protective immunity than the toxigenic vaccine strains [7], it was still good enough to save all the sheep on that fateful day at Pouilly-le-Fort. (The modest immunogenicity of Pasteur's vaccine may well have been caused by the persistence of a small proportion of pX01⁺/pX02⁺ cells, which thus remained toxigenic and fully virulent.)

The genes encoding the capsule and each of the three toxin components have been cloned in Escherichia coli [8-11], and the base sequences have been determined [12-14]. The protective antigen gene has been cloned in Bacillus subtilis, and immunization with the live recombinant strain protects guinea pigs from lethal challenge with virulent Bacillus anthracis spores [15].

None of the currently available anthrax vaccines is ideal, and efforts to develop better vaccines, with the use of newer molecular methods, represent an active area of research [16, 17]. Although the Sterne vaccine is effective and safe for use in many domestic animals (cattle, sheep, pigs, camels, buffaloes, and elephants), progressive disease caused by the vaccine strain has been observed in goats and llamas. The duration of protection conferred by the Sterne vaccine is somewhat limited, and the necessity for administration by injection is a disadvantage in many developing countries. The vaccine used in humans in the United States produces high titers of antibody to protective antigen in guinea pigs, but only animals vaccinated with the Sterne strain are completely protected against challenge with highly virulent strains of Bacillus anthracis.

Several different approaches are being used in the development of improved anthrax vaccines for humans. Purified protective antigen vaccines are being combined with adjuvants derived from the cell wall of the BCG strain of the tubercle bacillus [18] or with killed cells of Bordetella pertussis [19] to enhance the cellular immune response to the protective antigen. The Bacillus subtilis strain into which the protective antigen gene has been cloned, mentioned above [15], is a recombinant vaccine that does not contain the Bacillus anthracis genome. Transposon-induced mutagenesis has been used to produce mutant vaccine strains that cannot synthesize essential aromatic amino acids unavailable from the mammalian host and that thus cannot replicate for more than a few cycles in the mammalian host [16, 20]. Oral vaccines might consist of immunizing antigens from Bacillus anthracis cloned into appropriate bacterial vectors such as Salmonella species [17].

Anthrax has been known in humans and their animals for more than 7000 years, and the association between Bacillus anthracis and its hosts probably existed for millennia before it was recognized and recorded. In many countries, vaccination of animals and humans and improved methods of rearing livestock have effectively controlled the disease. Recent advances in molecular biology have allowed the elucidation of the precise mechanisms of virulence in Bacillus anthracis and give promise of even more effective vaccines. But it is difficult to imagine that the vast reservoir of Bacillus anthracis in the soil can ever be eradicated. The anthrax bacillus will surely endure; its place on our planet seems at least as secure as our own.