The mechanism by which anthrax kills its host remains unknown. Analyses of blood cultures of infected rabbits [1] and guinea pigs [2] indicate that the levels of bacteria in the blood rise steadily until the final hours of life, when they undergo a marked increase to reach levels of 107 and 109 organisms per ml of blood, respectively. This, coupled with repeated failures to identify lethal endotoxins or exotoxins associated with the bacteria or the fluids of an infected animal [3], led early workers to hypothesize that death was a consequence of mechanical obstruction caused by such a large number of organisms. However, some evidence suggested B. anthracis may produce a toxin. For example, subcutaneous injections of sterile crude extracts from skin lesions resulted in the production of similar lesions [4], and crude extracts of bacteria contained ”aggressions”, so called because, when mixed with sublethal doses of spores, they could cause death, and in vitro they were found to inhibit phagocytotic activities of polymorphonuclear leukocytes [3]. In addition, Bloom et al. [1] found in their analysis of blood cultures of infected rabbits that approximately 15% of injected animals had less than 103 bacteria per ml at the time of death.
This enigma led Smith & Keppie [2] to reexamine the question of whether B. anthracis causes death through over proliferation. When they measured the number of bacteria in the blood of infected guinea pigs they found that it increased from approximately 106 to 109 bacteria per ml in the 12 h preceding death. By treating infected guinea pigs with streptomycin they were able to prevent bacteremia and if the antibiotic was administered before the number of bacteria increased beyond 107 per ml, then the animals could be saved. Interestingly, if streptomycin was administered after this critical threshold, then death invariably was the outcome. These results clearly indicated that bacteremia was essential for death, but they also showed that the fate of the infected animals was determined when the number of bacteria present in the blood was two orders of magnitude lower than at the time of death (a point which will be returned to later). Thus, mechanical blockage could not adequately explain why infected animals died. Rather, a chain of events set in motion at an earlier stage of the infection led to death independent of bacterial proliferation.
Smith and Keppie [2] then went on to demonstrate that sterile-filtered serum derived from infected guinea pigs could induce edema when injected sub-cutaneously, or death when injected intravenously. Thus, the deadly effects of anthrax were in fact mediated by an exotoxin. They attributed previous failures to demonstrate the existence of an exotoxin to the higher concentration of toxin present in plasma versus peritoneal exudates, which they had previously used as a source of the toxin, and to the greater effectiveness of intravenous as opposed to intraperitoneal injection of the toxin.
A preliminary characterization of anthrax toxin suggested that it consisted of two components. Smith et al. [5] found that if they ultracentrifuged plasma from infected guinea pigs, the lethal activity of the resulting fractions was diminished relative to the original plasma preparation. However, if the supernatant was mixed with the pellet, then the toxic activity was restored. The pellet and the supernatant, therefore, contained two components, named factors I and II respectively, which synergistically united to form anthrax toxin.
In an elegant series of experiments, Stanley and Smith [6] found that while crude preparations of factor I plus purified factor II or purified factor I plus crude factor II caused skin edema in rabbits and were toxic when injected into mice, purified factor I plus purified factor II could cause skin edema but was not lethal. Thus, a third factor, called Factor III, was being lost during purification. In fact, factor III was found to be retained on a chromatographic column during purification of factor I, as well as in fractions discarded during the purification of factor II. Factor III alone or in combination with factor I was shown to be nontoxic. However, when combined with factor II, it was toxic for mice but produced no skin edema in rabbits. Preparations consisting of all three factors were synergistic for toxicity but produced edema of decreased size, indicating that factor III could competitively inhibit factor I in the production of edema. Shortly thereafter, similar results were obtained by Beall et al. [7] using different methods of purification.
Although these three factors have come to be equated with anthrax toxins, it is noteworthy that toxin produced in vivo differs from that produced and purified from cultures of B. anthracis in vitro in that it causes a more rapid death [8]. Thus, additional factors produced during the course of infection may enhance the lethal effects of anthrax toxin.
In the years following this work, the previous terminology was supplanted by another to reflect the activities of these factors; factor I became known as edema factor (EF); factor II, protective antigen (PA); and factor III, lethal factor (LF). More recently, the combination of EF plus PA has been referred to as edema toxin (EdTx) and that of LF plus PA as lethal toxin (LeTx) [9].
