Bacillus anthracis lethal toxin (PALF) stimulated the proliferation of human peripheral blood T-cells in vitro. Activation of T-lymphocytes by PALF required the presence of monocytes and did not result from a collaborative effect between T-cells and B-cells. PALF acted directly on monocytes and independently of T-cells. The monocytes contributed to the proliferation of T-cells by secretion of mediator(s). The mitogenic activity of the lethal toxin was dependent on its metalloprotease activity.
Bacillus anthracis, a spore-forming Gram-positive bacterium and the causative agent of anthrax, produces two exotoxins: the edema toxin (PAEF) and the lethal toxin (PALF) [1, 2]. These toxins and an antiphagocytic poly-d-glutamic acid capsule  are the known major virulence factors of this organism. Each of these toxins is composed of two proteins. The first is an enzyme, edema factor (EF, 89 kDa) or lethal factor (LF, 90 kDa), which is not able to cross the plasma membrane of the target cell. The second is a protein, protective antigen (PA, 83 kDa), involved in the internalization of these toxins. Anthrax lethal toxin has been identified to play a major role in the pathogenicity of B. anthracis infections [1, 4, 5]. In vitro this component is cytolytic for primary macrophages (Mφs) and Mφ cell lines such as J774A.1 and RAW264.7 , and induces death in experimental animals . Anthrax lethal toxin could be responsible for a wide range of effects including shock-like symptoms in humans and other species . The PALF intoxication process begins by the high affinity binding of PA to a cellular receptor . This binding is followed by a specific proteolytic cleavage of PA [9–11]. The release of the N-terminal 20-kDa fragment of PA exposes a high affinity binding site for either EF or LF [9, 12, 13]. Finally, the lethal toxin is internalized by receptor-mediated endocytosis through an acidic endocytic vesicle  and probably translocated into the cytosol . The exact process by which the anthrax lethal protein causes cell death is still poorly understood. A zinc metalloprotease consensus sequence (H686EXXH690), present in the sequence of LF, is required for in vitro lethal toxin activity . Thus lethal toxin was proposed as a Zn2+-metalloprotease acting on an intracellular substrate. However, despite intensive investigations, the specific target(s) of PALF remain obscure. Lethal toxin, at subcytolytic concentrations, causes the release of tumor necrosis factor (TNF-α) and interleukin-1β (IL-1β) from macrophage-like cells, RAW264.7 . Furthermore, macrophages may generate reactive oxygen intermediates (ROIs) in response to PALF challenge and adding exogenous antioxidants such as β-mercaptoethanol to macrophages protected cells from the lethal effects of the toxin . Alternatively, lymphocyte mitogenesis agents such as the potent tumor promoter phorbol myristate acetate (PMA) induces oxygen radicals in macrophages and neutrophils [17, 18]. Moreover, PMA may initiate lymphocyte proliferation by its potent generation of oxygen radicals . Therefore it was crucial to investigate the possible effect of PALF as an immunomodulator. Bacterial exotoxins can exert their immunomodulatory effects on various lymphoid cell populations. Here, we investigated whether or not PALF, as a stimulator of intracellular ROIs, could initiate human lymphocyte proliferation.
Materials and methods
Human mononuclear cells were separated from heparinized (50 units ml−1) peripheral blood of human healthy donors by Ficoll-Paque (Pharmacia Biotech, Uppsala, Sweden) gradient centrifugation . Cells were resuspended in RPMI 1640 (Gibco BRL, Gaithersburg, MD) supplemented with 2 mM l-glutamine, 5% of fetal bovine serum, streptomycin (100 μg ml−1), and penicillin (100 IU ml−1) at 37°C in a humidified atmosphere of 95% air and 5% CO2. Highly enriched human B-lymphocytes were obtained by immunomagnetic isolation from mononuclear cells, using Dynabeads M-450 Pan-B - CD19 (Dynal A.S., Oslo, Norway). Monocytes were obtained from the remaining T-cell-monocyte population by two successive incubations in tissue culture flasks (TPP, Switzerland) at a concentration of 5×106 human mononuclear cells per ml of RPMI 1640 medium for 90 min at 37°C. Monocytes were obtained by mechanically scraping the adherent cells (>98% pure as assessed by morphology).
PA and LF components from B. anthracis strains were purified by chromatography on a Q Fast flow column (FPLC, Pharmacia, Uppsala, Sweden) using a gradient of 0–1 M NaCl in 20 mM Tris-HCl at pH 7.5. Toxic shock syndrome toxin 1 (TSST-1) was a gift of S. Giudicelli and staphylococcal enterotoxin B (SEB) was a gift of Dr. I. Motta (Institut Pasteur, France).
Proliferative responses were monitored by DNA synthesis assays. Briefly, cell density was adjusted to 5×106 human mononuclear cells per ml, and 100 μl of the cell suspension was transferred to flat-bottomed 96-well plates (Falcon) and incubated with 30 ng ml−1 of PA, or/and LF, LF686 (alanine for histidine in position 686) at 37°C in a humidified atmosphere of 95% air and 5% CO2. Proliferative responses were assessed by [3H]thymidine uptake (1 μCi ml−1; specific activity, 40–60 Ci mmol−1; Amersham). Cells were labeled with [3H]thymidine 18 h before harvesting.
Results and discussion
PALF primes human lymphocyte proliferative responses
To assess the mitogenic potential of the anthrax lethal toxin, we incubated peripheral blood mononuclear cells (PBMC), as an ex vivo model, with various concentrations of the purified proteins, PA and LF components from B. anthracis strains. Human mononuclear cells were isolated as described in Section 2. The viability of cells assessed by the trypan blue exclusion test was >98%. 5×105 human mononuclear cells were incubated with PALF (30 ng of each protein) and on day 5 the [3H]thymidine incorporation was measured (Fig. 1). The PALF complex induced a dramatic increase of the [3H]thymidine incorporation. Thymidine uptake correlated with morphological changes in the lymphocytes, switching from a spherical to a polarized shape (Fig. 1, inset), and demonstrated T-cell activation. With either PA or LF, we detected no significant increase of the [3H]thymidine incorporation. Thus, the proliferation of human mononuclear cells was PALF-dependent. PALF was mitogenic for human mononuclear cells; the magnitude of the response ranged from 4000 to 12 000 cpm depending on the donor. The maximal proliferation response was obtained between days 4 and 6 with an optimal dose of 0.01 μg PALF per 5×105 human mononuclear cells (data not shown). Moreover, this result indicated a possible role of the activity of the lethal toxin in the proliferation phenomenon. To address whether the metalloprotease activity of the lethal factor is required for this phenomenon, we tested the proliferation of PBMC with fully inactivated LF. We used the fully inactivated LF protein obtained by site-directed mutagenesis that substituted alanine for the histidine in position 686 (LF686) . As expected, the macrophage cell lines J774A.1 and RAW264.7 were resistant to PALF686 at concentrations as high as 10 μg ml−1 (50% cytotoxicity of native PALF occurred at 35 ng ml−1, data not shown). With PALF686, in contrast to native PALF, we observed no proliferation of human lymphocytes (Fig. 1). Therefore the mitogenic activity mediated by PALF required the native zinc metalloprotease consensus sequence of LF and thus full metalloprotease activity of the protein.
Monocyte dependence of the mitogenic response
We investigated whether the mitogenic response was dependent on the presence of monocytes or resulted from a collaborative effect between T- and B-cells. Thus we analyzed the proliferation of fractionated lymphocyte populations stimulated with PALF. Efficiencies of the cell purification were assessed by the well known proliferative capacity of macromolecules acting as microbial superantigens  including Staphylococcus aureus protein TSST-1 [22–24] and SEB . Fractionated lymphocyte populations were preincubated with PALF, SEB or TSST-1 and [3H]thymidine incorporation was measured (Fig. 2). After preincubation with TSST-1 or SEB, we observed no proliferation of B-cells co-cultured with monocytes (Fig. 2; lane E) or T-cells (Fig. 2; lane C). In contrast, monocytes plus T-cells (Fig. 2; lane D) were stimulated by either toxin. Therefore TSST-1 (500 ng ml−1) and SEB (50 ng ml−1), here used as controls, caused the expected T-cell monocyte-dependent proliferative response. This result confirmed the purity of each cell population. The viability of the isolated cell population(s) implicated in proliferation via PALF was verified with a reconstituted population (Fig. 2; lane B). The proliferative capacities between the reconstituted population (Fig. 2; lane B) and the original population (Fig. 2; lane A) were not significantly different. Neither population involved in the mitogenic phenomenon mediated by PALF was inactivated during cell isolation. However, with PALF, B-cells did not react even in the presence of T-cells (Fig. 2; lane C) and also excluded a collaborative effect between T-cells and B-cells. In contrast, PALF elicited a significant proliferation of T-cells co-cultured with monocytes (Fig. 2; lane D). The removal of adherent cells eliminated the incorporation of [3H]thymidine. Therefore, the effective activation by PALF of the responding T-cell population is clearly monocyte-dependent.
Role of monocytes
Having established that the mitogenic effect of PALF was monocyte-dependent, subsequent work was focused mainly on the contribution of human monocytes to this phenomenon. To control the direct effects of monocyte products on the T-cells via PALF we used filtered conditioned medium from monocytes (0.22 μm; Millex-GV4, Millipore) incubated with PALF [f(M+PALF)] or without PALF [f(M)]. As a control, filtered conditioned media obtained after incubation of monocytes with TSST-1 [f(M+TSST-1)] or SEB [f(M+SEB)] were tested in the presence of T-cells. The [3H]thymidine uptake of T-cells with f(M+TSST-1) or f(M+SEB) corresponded respectively to 350 cpm or 270 cpm. This result demonstrated the absence of contamination of the filtered medium by monocytes. Fig. 3 shows the ability of the different filtered media to stimulate T-cells incubated with PALF (0.03 μg of each protein). The f(M+PALF) but not f(M) allowed the proliferation of T-cells incubated with PALF. Thus we conclude that no T-cell-monocyte contact was required for T-cell proliferation via PALF. The role of monocytes in proliferation via PALF is not to provide accessory functions but rather to provide mediators. Also, the production of mediator(s) by monocytes was primed by native PALF. We performed additional experiments to determine if this effect is metalloprotease activity-dependent by using the fully inactivated LF protein (LF686). The filtered medium f(M+PALF686), in contrast to f(M+PALF), did not stimulate the proliferation of T-cells incubated with PALF (Fig. 3). Therefore, the metalloprotease activity of PALF is necessary for the production and release of inflammatory mediator(s) by monocytes and the subsequent T-cell proliferation.
The data reported herein strongly suggest that the anthrax lethal toxin is a T-cell mitogen. The simplest model that might account for this mitogenesis scenario is to propose that PALF acts directly on human monocytes independently of T-cells. Upon exposure to PALF, human monocytes, the primary effector cells in inflammation, are primed via the PALF metalloprotease activity for the release of one (or more) mediator(s). Finally, a direct interaction with T-cells and released monocyte-mediator(s) triggers and activates T-cells causing proliferation. The in vitro PALF mitogenic properties on human T-cells that we described here may modulate the host defense system and may cause profound in vivo alterations in the immune system homeostasis. The lethal toxin may induce inflammation thereby causing a massive and uncontrolled stimulation leading to lethal shock. The activity of PALF described here is likely a crucial point in the pathogenesis of anthrax. We are investigating the molecular basis and the identity of the mediator(s) mainly responsible of the T-cell stimulation by B. anthracis PALF in humans.
We are greatly indebted to I. Motta for the generous gift of SEB, to S. Giudicelli for the generous gift of TSST-1.