Macrophages Are Sensitive to Anthrax Lethal Toxin through an Acid-dependent Process*

Anthrax lethal toxin, which consists of two proteins, protective antigen and lethal factor, is lethal for experimental animals. This study describes the first in vitro system demonstrating lethality of the toxin. Mouse peritoneal macrophages are killed within 1 h of exposure to the toxin. Neither protein component alone shows any toxic activity. The minimal effective concentration of protective antigen and lethal factor was and hg/ml, respectively. None of the sev- eral established cell lines examined was killed. Cells could be completely protected from the toxin by pre- treatment with agents, such as amines or monensin, which dissipate intracellular proton gradients and raise the pH of intracellular vesicles. This protection was reversible and could be overcome by lowering the intravesicular pH. Antitoxin added after preincubation with amines was unable to protect cells subsequently exposed to low pH treatment. These results suggest that anthrax lethal toxin requires passage through an acidic endocytic vesicle in order to exert its toxic effect within the cytosol. cell monolayers. Lactic dehydrogenase was measured in cells lysed in 100 pl of 0.05% digitonin, by oxidation of lactate to pyruvate in the presence of nicotinamide adenine dinucleotide (14), using an automated analyzer (COBAS, Roche Analytical Instruments, Inc., Nutley, NJ). All the lactic dehydrogenase released from toxin-treated monolayers was recovered in the medium and was nonsedimentable at 500 X g. Thus, loss of lactic dehydrogenase from the monolayer represents cell lysis. Results are expressed as the per cent of control cellular lactic dehy- drogenase f S.E. in triplicate cultures unless otherwise indicated. Toxicity was also measured by exclusion of trypan blue (0.05 g/lOO ml) from cells incubated in phosphate-buffered saline. The established cell lines, L929 and 3T3, were obtained from the American Type Culture Collection (Rockville, MD). MRC-5 and FRL-103 cells were from The Salk Institute (Swiftwater, PA).

Anthrax lethal toxin, which consists of two proteins, protective antigen and lethal factor, is lethal for experimental animals. This study describes the first in vitro system demonstrating lethality of the toxin. Mouse peritoneal macrophages are killed within 1 h of exposure to the toxin. Neither protein component alone shows any toxic activity. The minimal effective concentration of protective antigen and lethal factor was and hg/ml, respectively. None of the several established cell lines examined was killed. Cells could be completely protected from the toxin by pretreatment with agents, such as amines or monensin, which dissipate intracellular proton gradients and raise the pH of intracellular vesicles. This protection was reversible and could be overcome by lowering the intravesicular pH. Antitoxin added after preincubation with amines was unable to protect cells subsequently exposed to low pH treatment. These results suggest that anthrax lethal toxin requires passage through an acidic endocytic vesicle in order to exert its toxic effect within the cytosol.
The toxic nature of anthrax infection was suspected by Koch (1) in his earliest studies of its pathogenesis. Later studies by Bail and Weil (2) and others (3) using extracts of infected tissue suggested that aggressins or toxins were important factors in anthrax virulence. Subsequently, a toxin was demonstrated definitively by Smith and co-workers (4). Further research has established the existence of two toxin complexes (5, 6). The edema toxin consists of two proteins: protective antigen, or Factor 11, together with edema factor, or Factor I. Each protein has an apparent molecular weight of approximately 85,000-90,000 (7). This toxin produces edema in experimental animals (8,9) and edema factor has recently been shown to be a calmodulin-dependent adenylate cyclase (10). The lethal toxin consists of protective antigen plus a third protein, lethal factor, or Factor 111, with an apparent molecular weight of 83,000 (7). It is this toxin that is lethal for several animal species (8, 9). The only biological system presently available for studying lethal toxin is lethality in experimental animals. The mechanism of action of the toxin remains unknown. In this communication, I report the first in vitro system demonstrating the toxicity of anthrax * The views of the author do not purport to reflect the positions of the Department of the Army or the Department of Defense. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked ''advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. lethal toxin and show that the expression of toxin activity requires an acidic intracellular environment.

MATERIALS AND METHODS
Peritoneal exudate macrophages were obtained from male C3H/ HeNHsd mice (Harlan/Sprague Dawley, Inc., Walkersville, M D ) as previously described (12). They were plated at 7.5 X lo5 cells in 1 ml in a 2-cmz 24-well tissue culture plate unless otherwise noted. After 1-2 h, the nonadherent cells were removed by washing, and complete medium, consisting of Dulbecco's modified Eagle's medium containing 20% fetal bovine serum, penicillin (100 units/ml), streptomycin (100 pglml), and 10% L cell conditioned medium (11) was added. Protective antigen, lethal factor, edema factor, and a goat antiserum to lethal factor were generously supplied by S. Leppla (this Institute).
The protective antigen, lethal factor, and edema factor toxin components were purified by previously published methods to at least 90% homogeneity when analyzed by sodium dodecyl sulfate gels (13). Cell toxicity was determined by the amount of cytoplasmic lactic dehydrogenase present in control or toxin-treated cell monolayers. Lactic dehydrogenase was measured in cells lysed in 100 pl of 0.05% digitonin, by oxidation of lactate to pyruvate in the presence of nicotinamide adenine dinucleotide (14), using an automated analyzer (COBAS, Roche Analytical Instruments, Inc., Nutley, NJ). All the lactic dehydrogenase released from toxin-treated monolayers was recovered in the medium and was nonsedimentable at 500 X g. Thus, loss of lactic dehydrogenase from the monolayer represents cell lysis. Results are expressed as the per cent of control cellular lactic dehydrogenase f S.E. in triplicate cultures unless otherwise indicated. Toxicity was also measured by exclusion of trypan blue (0.05 g/lOO ml) from cells incubated in phosphate-buffered saline. The established cell lines, L929 and 3T3, were obtained from the American Type Culture Collection (Rockville, MD). MRC-5 and FRL-103 cells were from The Salk Institute (Swiftwater, PA).

RESULTS AND DISCUSSION
Initial experiments clearly indicated that the combination of protective antigen and lethal factor, each at 1 pg/ml, was cytotoxic for mouse peritoneal macrophages and caused destruction of the cell monolayer after a 24-h exposure. One hundred per cent of the toxin-treated cells were dead as determined by trypan blue exclusion while untreated control cells remained alive. Protective antigen or lethal factor alone caused no morphological change in the macrophages and the cells were all viable. Quantitation of the cytotoxicity by release of cytoplasmic lactic dehydrogenase (Fig. l), revealed the minimal effective concentration of lethal factor to be pg/ml when protective antigen was present at 1 pg/ml. The minimal effective concentration of protective antigen was =lo-' pg/ml when the lethal factor concentration was maximal (1 pg/ml). Protective antigen or lethal factor alone at 1 pg/ml was completely nontoxic. Identical results were obtained by using trypan blue exclusion as the measure of cytotoxicity. Five different preparations of protective antigen and lethal factor showed similar potencies. Cell death was first detectable after 1-h exposure to protective antigen + lethal factor (each at 1 pg/ml) and was complete by 4 h. Edema toxin (protective antigen and edema factor, each at 1 pg/ml) caused no toxicity after 4 h.
Previous studies of the effect of anthrax toxins on cells in vitro performed as early as 1911 showed that extracts of tissue from anthrax-infected animals inhibited leukocyte killing of anthrax bacilli (2). Subsequent workers reported that anthrax culture filtrates inhibited neutrophil chemotaxis (15)  phages were cultured, exposed to varying concentrations of protective antigen and lethal factor, and assessed for toxicity 24 h later by the amount of lactic dehydrogenase remaining in the cell monolayer as described in the text. The per cent of control cellular lactic dehydrogenase is plotted uersus the concentration of lethal factor (log scale). Protective antigen was used at varying concentrations as indicated. PA, protective antigen; LDH, lactic dehydrogenase.
tive antigen + edema factor + lethal factor) inhibited neutrophil phagocytosis (16). No cytopathic effect was noted and individual components were not tested. These previously observed effects may reflect the ability of the edema toxin (protective antigen + edema factor) present to raise the cyclic AMP level of cells (10) and inhibit their function (17). Anthrax culture filtrates produced no cytopathology in KB cells, primary mouse embryo cells, or a guinea pig spleen cell line (18). Similarly, Fedotova (19) noted no cytopathology of anthrax toxin for F1 or HeLa cells. However, he noted some growth inhibition in these cells and degenerative changes in guinea pig macrophages which were evident at 4 h but not at later times. Based on the present results, it is likely that these morphological effects were caused by the protective antigen + lethal factor in the crude preparation. However, no quantitation was reported and purified components were not used. The difference between the lack of cytotoxicity reported for various cell lines and the present results with macrophages may be due to differences in sensitivity to lethal toxin, although individual toxin components were not tested in the previous studies. Using purified protective antigen + lethal factor, each at 1 pg/ml, we found no significant cytotoxicity after 24-h exposure (~2 0 % lethality) for four cell lines (MRC-5, L929, FRL-103, and 3T3) (data not shown). The basis for this marked difference in sensitivity between macrophages and other cells awaits further study.
In view of the large amount of evidence showing that some toxins (20,21), other protein ligands (22), and viruses (23,24) enter and exert effects on cells by an acid-dependent process, we next examined the effects on lethal toxin activity of agents which dissipate intracellular proton gradients. Preincubation of macrophages with the lysomotropic amines, NH4C1 and chloroquine, or with the ionophore monensin, for 20 min before exposure to protective antigen + lethal factor, completely protected them from killing by lethal toxin (Fig. 2). Other experiments demonstrated that the drugs themselves were nontoxic. The protection by NH4C1 was totally reversible by removing the NH&l along with the toxin 1 h after exposure to protective antigen + lethal factor (data not shown). These results strongly suggest that anthrax lethal toxin exerts its Protective antigen + lethal factor, each at 0.1 pg/ml, was then added and toxicity was measured 4 h later as described in the text. Inhibitor was present throughout the experiment. The per cent control cellular lactic dehydrogenase is plotted uersus the concentration of inhibitor. A, cells exposed to inhibitor plus protective antigen + lethal factor; 0, cells exposed to inhibitor alone. LDH, lactic dehydrogenase.
effect within the cytosol after passage through an acidic intracellular vesicle or compartment. If the protection by amines is due to their ability to raise the pH of intracellular vesicles, then it might be possible to overcome such protection by lowering the pH as has been previously demonstrated with diphtheria toxin (25,26), epidermal growth factor (22), and some viruses (23). Macrophages protected from lethal toxin by preincubation with NH&l were exposed for 10 min to media of varying pH in the presence of NH4Cl to reduce the pH of intracellular vesicles (27). They were then placed in toxin-free medium (pH 7.4) with NH4Cl and assayed for toxicity 24 h later. NH4C1 was present throughout the experiment. Exposure to media of increasingly lower pH overcame the inhibition of toxicity by NH&l (Fig. 3). Thus, at pH 54.75, complete  FIG. 3. Reversal of the NH4C1 inhibition of protective antigen + lethal factor toxicity by low pH. Macrophages were preincubated with medium containing 10 mM NH&l for 30 min. Protective antigen (1 pg/ml) and lethal factor (0.1 pg/ml) were then added in the presence of 10 mM NEC1 and cells were incubated for another 90 min. Cells were then washed and reincubated with medium with 10 mM NH4Cl and no toxin for an additional 60 min. At that time they were exposed for 10 min to medium with 10 mM NH&l and 20 mM HEPES' adjusted to different pH with HC1. The cells were then washed again and reincubated in medium with 10 mM NH&l (pH 7.4). Toxicity was measured 24 h later and expressed as the cellular lactic dehydrogenase activity (milliunits/monolayer) of triplicate wells +. S.E. versus pH of the medium during the 10-min exposure.
A, toxin-treated cells; 0, control cells. NH&l was present thoughout the entire experiment. Cells not exposed to NH4Cl that were given protective antigen + lethal factor and treated at pH 7.2 had <5 milliunits/monolayer. LDH, lactic dehydrogenase.  toxicity of anthrax lethal toxin was restored. Similarly, it was also possible to overcome the inhibition of toxicity afforded by monensin by treating cells for 10 min at pH 4.5 (data not shown). These results suggest that the protection by amines and monensin is due to the elevated vesicular pH and support the conclusion that a low pH is required for anthrax lethal toxin activity.
To determine the cellular location of the toxin after pretreatment with NH4C1 at the time when low pH exposure can reverse the inhibition of toxicity, we studied whether anti-' The abbreviation used is: HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid.
As described in the legend to Fig. 3, macrophages were preincubated with 10 D M NH4C1 for 30 min then exposed to protective antigen (1 pg/ml) plus lethal factor (0.1 pg/ml) for 90 min, washed, and reincubated in medium with 10 mM NH4Cl without toxin for a further 60 min. At this time, cells were washed three times with cold medium and incubated with or without antitoxin (1:50 dilution of antiserum) in the presence of 10 mM NH4Cl for 60 min at 4 "C. They were then washed 3 times and exposed for 10 min at 37 "C to medium at pH 4.5 as described in Fig. 3. After being washed again they were reincubated in medium with 10 mM NH&l (pH 7.4) and assayed for toxicity 24 h later. N&Cl was present at all times throughout the experiment.
Results represent the mean f S.E. of triplicate wells. We first showed that antibody to lethal factor, when added at 4 "C, was able to block the effect of protective antigen + lethal factor, located on the cell surface after adsorption at 4 "C (Table I). Additional controls showed that this protection was not affected by including 10 mM NH&l during the antibody incubation or by treating cells at pH 4.5 for 10 min after the antibody exposure. The next experiment showed that antitoxin, added at 4 "C, after the NH4C1 block and immediately before the acid pH treatment, was unable to prevent the acid-induced toxicity ( Table 11). A further control showed that the amount of antitoxin used in the experiment was more than sufficient to neutralize all of the lethal factor initially added (data not shown). This result demonstrates that after pretreatment with NH4C1, the toxin is not accessible to antitoxin and supports the idea that it is present in an intracellular location before expressing its toxic effect on the cell. This interpretation is consistent with previous reports suggesting that amines and monensin cause accumulation of diphtheria toxin (25, 26, 28) and several protein ligands (29) within acidic intracellular vesicles. The protection against toxicity afforded by lysomotropic amines and monensin, the reversal of this protection by low pH treatment, and the inability of antitoxin to neutralize toxicity after amine treatment all suggest that anthrax lethal toxin must pass through an acidic endocytic vesicle to exert its toxic effect within the cytosol. The identification of the macrophage as a cell sensitive to anthrax lethal toxin provides an in vitro system which is necessary to study further the mechanism of toxin action.