1. Introduction
Humans acquire anthrax by exposure to
Bacillus anthracis spores through cutaneous, gastrointestinal, inhalational, or blood-borne routes. High levels of circulating bacteria occur in systemic anthrax [
1]. Baboons infused intravenously with vegetative
B. anthracis bacilli mimic the systemic disease as exhibited by key features of sepsis [
2,
3], a life threatening, dysregulated immune response to infection that results in organ failure and often leads to death. Bacterial sepsis is associated with high levels of lymphocyte apoptosis [
4,
5] and increased levels of circulating nucleosomes [
6,
7] that may arise from uncleared apoptotic cells that have become secondarily necrotic [
8]. Nucleosomes contribute to acute septic pathology by promoting intra-alveolar hemorrhage, macro- and microvascular thrombosis, and organ dysfunction [
9].
Lymphoid organ macrophages are responsible for the clearance of sudden increases in apoptotic cells by a process known as efferocytosis [
10,
11]. The inhibition of efferocytosis in macrophages may exacerbate sepsis by increasing the burden of sepsis-promoting histones and other damage-associated molecular patterns secondary to defective apoptotic cell clearance. Efferocytosis has been reported to be inhibited by elevated cellular cAMP [
12] and requires the binding of macrophages to apoptotic cells followed by macrophage signaling events that lead to Rac1-dependent apoptotic cell engulfment [
13,
14,
15]. Direct binding is mediated by tethering receptors, while indirect binding occurs via soluble proteins that bridge the binding of apoptotic cells to macrophages. There are approximately 12 known signaling receptors that can be divided into (i) those that require bridge proteins to bind apoptotic cells and (ii) those that do not [
13]. Among the former, which were evaluated in this study, Tyro3, Axl, and MerTK (TAM family) require the bridge proteins Gas6 or Protein S [
16], while αVβ3 and αVβ5 require MFGE8 [
17,
18] or CCN1 [
19]. Efferocytic macrophages in secondary lymphoid organs express MerTK [
10] and alternative/M2-like markers CD163 and CD206 [
20]. Glucocorticoids such as dexamethasone (Dex), which have been historically used to treat severe sepsis [
21,
22], enhance macrophage efferocytosis by increasing the expression of the efferocytosis receptor MerTK and its cofactors Protein S and Gas6 [
23,
24].
In addition to its poly-
d-glutamic acid capsule, the major known
B. anthracis virulence factors include Lethal Toxin (LT) and Edema Toxin (ET), formed by the association of the cell-binding protein Protective Antigen (PA) with the active components Lethal Factor (LF) or Edema Factor (EF), respectively [
25,
26]. PA binds to at least two independent receptors on the target cells [
27,
28], undergoes cleavage and multimerization on the cell surface, and then facilitates the binding and translocation of the LF/EF moieties into the cytosol, where they exert their toxic activities. EF is a calcium- and calmodulin-dependent adenylate cyclase that increases intracellular cAMP concentrations to supraphysiologic levels [
29]. Mechanisms of ET-induced virulence and tissue damage during infection are not fully understood but may involve the inhibition of innate immunity during early stage infection and direct effects on liver tissue [
26]. ET has numerous effects on immune cells, such as the inhibition of macrophage chemotaxis [
30] and phagocytosis [
31], the rescue of macrophages from Toll-like receptor 4-induced apoptosis [
32], the inhibition of neutrophil priming and motility [
33,
34,
35], the alteration of dendritic cell cytokine secretion, maturation and chemotaxis [
36,
37,
38], the suppression of T cell activation and chemotaxis [
30,
39,
40], and the skewing of CD4
+ T cell differentiation to the Th2 subset [
41].
As efferocytosis is sensitive to cAMP [
12], this study tested the hypothesis that ET inhibits efferocytosis initiated by MerTK and integrin αVβ5 signaling pathways and explored the intracellular signaling events impacted. The results demonstrate that ET inhibits macrophage-mediated efferocytosis, Rac1 signaling, and the phosphorylation of Ca
2+/calmodulin-dependent protein kinases, Rac1 and vasodilator-stimulated phosphoprotein (VASP) induced by apoptotic cell exposure.
3. Discussion
Edema Toxin is a recognized virulence factor of
B. anthracis that functions as a highly active calmodulin-dependent adenylate cyclase, raising the concentration of cAMP in cells to supraphysiologic levels [
29]. A selective but unidentified role for ET in systemic anthrax infection was recently demonstrated by the reduced survival of macaques infected intravenously with mutant bacilli lacking LF (ET function intact) compared to macaques similarly infected with mutant bacilli lacking PA (neither toxin present) [
59]. Although ET was previously shown to inhibit the phagocytosis of
B. anthracis bacilli through altered cytoskeletal remodeling [
31], defective bacterial clearance did not appear to explain the recently identified selective role of ET in late stage anthrax in the macaque intravenous infection model [
59]. Moreover, the specific effects of ET on macrophage cytoskeletal remodeling have not been previously defined beyond the roles identified for both Protein Kinase A and exchange protein activated by cAMP (EPAC), which act immediately downstream of cAMP [
31].
In the present study, we tested the hypothesis that ET inhibits efferocytosis, a process of apoptotic cell engulfment and clearance that depends on cytoskeletal remodeling requiring Rac1 activation [
48] and that has been shown to be inhibited by cAMP [
12]. As predicted, we observed that ET inhibited efferocytosis in a dose-dependent manner. The 50% effective concentration for the inhibition of efferocytosis was calculated to be 0.6 nM. Given that circulating LT levels (defined as PA63+LF) in late stage inhalational anthrax in Rhesus macaques can reach 1150–12,100 ng/mL, equivalent to 21–79 nM [
60] and that the ratio of LF:EF in the circulation of lethally infected rabbits is consistently 5:1 at various time points [
61], the capacity of ET to inhibit efferocytosis is within the predicted physiologic range of ET encountered during systemic anthrax.
What are the predicted pathologic consequences of impaired efferocytosis during systemic anthrax? Data from lethal human cases of inhalational anthrax and a baboon intravenous bacillus infection model are consistent with sepsis being a factor that contributes to mortality in systemic anthrax [
2,
3]. A key feature of human sepsis pathology is rapid and widespread lymphocyte apoptosis [
5]. Failure to effectively clear apoptotic lymphocytes leads to their secondary necrosis and the subsequent release of nucleosomes/histones and other toxic damage-associated molecular patterns [
62]. Circulating nucleosomes/histones cause key features of sepsis pathology, including intra-alveolar hemorrhage, macro- and microvascular thrombosis, and organ dysfunction [
9]. Based on these predictions, it was important that we test the impact of ET on efferocytic macrophages having the phenotype of those residing in human secondary lymphoid organs. Efferocytic macrophages in human lymph nodes express CD163 and CD206 [
10]. We showed that human monocyte-derived macrophages either expressed or could be induced to express these markers by polarization with IL-4+IL-10+Dex or Dex alone and that ET could inhibit efferocytosis of macrophages with this phenotype.
Understanding the precise mechanisms whereby ET inhibits efferocytosis could lead to new strategies to restore efferocytosis in the setting of late stage anthrax and other forms of acute sepsis. To produce therapeutic effects, however, such strategies should preserve bacterial phagocytosis and/or be used in the context of effective antibiotic therapy. Thus, to further understand how ET subverts intracellular signaling events required for the engulfment of apoptotic cells, efferocytosis assays were conducted under defined conditions. We used conditions that promoted the initiation of efferocytic signaling through MerTK, a relevant efferocytic signaling receptor in the secondary lymphoid organs [
10] or through αVβ5 integrin that can signal for efferocytosis on its own [
18] or in cooperation with MerTK [
63]. The expression of MerTK, Tyro 3, and αVβ5 (but not Axl or αVβ3) by M2(Dex) macrophages ensured that signaling initiated by Protein S- or MFGE8-opsonized apoptotic PMN preferentially occurred through these receptors.
An evolutionarily conserved signaling pathway that leads to efferocytosis involves the recruitment of a complex of p130Cas, CrkII, and DOCK-180 proteins to the vicinity of the signaling receptor, resulting in Rac1 activation that leads to cytoskeletal reorganization and engulfment [
64]. Experiments conducted in HEK-293T cells showed that MerTK activation leads to the Src-mediated phosphorylation of FAK on Tyr 861, followed by the recruitment of FAK to αVβ5, the increased phosphorylation of p130Cas, the increased formation of the p130Cas/CrkII/Dock180 complex, enhanced Rac1 activation, and efferocytosis [
63]. The signaling was enhanced synergistically by the presence of αVβ5 [
63]. Our studies in primary human macrophages revealed that ET could substantially and reproducibly inhibit the activation of Rac1 following the exposure of M2(Dex) macrophages to Protein S-opsonized apoptotic PMN, showing a perturbation of the efferocytic signaling pathway by ET. Although we did not observe a significant inhibition of the phosphorylation of Tyr 861 on FAK across three independent donors, a 49% and 28% reduction of phosphorylation of this residue in the presence of ET was observed in two of the three donors (not shown). We did, however, observe a modest but consistent reduction in phosphorylation of FAK Tyr 925. Notably, the phosphorylation of this residue was required for FAK-mediated cell protrusion in mouse embryonic fibroblasts that depended on the p130Cas/Dock180/Rac1 signaling pathway [
56], suggesting that an interruption of the FAK activation by ET may contribute to its capacity to inhibit efferocytosis. Furthermore, the phosphorylation of FAK at Tyr 925 played a causative role in stimulating the migration of RAW 264.7 mouse macrophages [
65], consistent with a role for the inhibition of FAK in the ET-mediated suppression of cell migration [
30,
35].
The most substantial changes in cytoskeleton signaling protein phosphorylation events downstream of MerTK that were induced by the pre-exposure of macrophages to ET included the reduced phosphorylation of Ser-71 on Rac, as well as the reduced phosphorylation of activating Thr residues of CamKs 1α and 4 and Ser residues of VASP. Though originally reported to be an inhibitory phosphorylation event, Ser-71 phosphorylation on Rac was more recently found to promote filopodia formation in fibroblasts [
58]. The predicted effects downstream of Cam kinases [
50,
51] include altered transcription. Notably, ET has been shown to alter the transcription of macrophages as early as 3 h posttreatment, and these changes have been proposed to impact cytoskeletal remodeling [
31]. VASP has not previously been implicated in efferocytosis, but its localization to the phagocytic cup was shown to be essential for actin cytoskeletal reorganization, the extension of pseudopodia, and phagocytosis by macrophages [
66]. In Drosophila macrophages, Ena/VASP stimulates WAVE regulatory complex-mediated actin assembly in the presence of Rac, which is necessary for lamellipodia formation [
67], an event also known to be essential for efferocytosis. Consistent with an impact of these phosphorylation events on filopodia formation, we observed a reduced filamentous actin-positive filopodia in ET-treated CD11b
+ M2(Dex) macrophages exposed to apoptotic PMN (
Figure 7). However, this observation requires confirmation in live cell imaging studies to effectively capture actin reorganization events occurring during the process of efferocytosis.
This study establishes the capacity of B. anthracis ET to inhibit macrophage-mediated efferocytosis at doses relevant to systemic infection and further provides novel insight into putative molecular mechanisms.
4. Materials and Methods
4.1. Isolation, Differentiation, and Polarization of Human Monocytes
The acquisition and use of all human cells in this study was compliant with the Declaration of Helsinki and was reviewed and approved by the Oklahoma Medical Research Foundation Institutional Review Board, ensuring the protection of human subjects. Fresh buffy coats (containing white blood cells and platelets) of healthy blood donors were purchased from the Oklahoma Blood Institute, USA. Mononuclear cells were isolated using Lympholyte-H Cell Separation Media (Cedarlane, Burlington, NC, USA). The monocytes were either purified using an EasySep Human Monocyte Isolation Kit (StemCell Technologies, Cambridge, MA, USA) or isolated as adherent cells according to the following procedure: isolated mononuclear cells were resuspended at 4 × 106/mL in Iscove’s Modified Dulbecco’s Medium (IMDM, Gibco, Grand Island, NY, USA) and adhered to nontreated 100 mm × 20 mm tissue culture dishes (Corning, Corning, NY, USA) for 1 h at 37 °C in 5% CO2. Nonadherent cells were removed by washing with IMDM. To generate macrophages, monocytes were cultured for 6 or 7 days with 50 ng/mL M-CSF (Peprotech, Rocky Hill, NJ, USA) in either RPMI 1640 with 10% FBS, 2 mM glutamine, 100 U pen/strep/mL, 10 mM HEPES buffer, and 1 mM sodium pyruvate (for M2(IL-4+IL-10+Dex) macrophages) or in IMDM containing 10% FBS and 100 U pen/strep/mL (for M2(Dex) macrophages). On day 7 or 8, adherent cells were detached using a Trypsin-EDTA solution (Millipore Sigma, St. Louis, MO, USA) and resuspended in a fresh medium with 50 ng/mL M-CSF and 20 ng/mL IL-10 (Peprotech), 20 ng/mL IL-4 (Peprotech), and/or 10 nM dexamethasone (Dex) (Millipore Sigma) at 5 × 105/well in temperature-sensitive 12 well-plates (Nunc UpCell, Rochester, NY, USA) overnight.
4.2. Generation of Early Apoptotic Human Polymorphonuclear Neutrophils (PMN)
Polymorphonuclear neutrophils (PMN) were isolated from fresh human blood using the EasySep Direct Human Neutrophil Isolation Kit (StemCell Technologies, Cambridge, MA, USA). The isolated PMN were labeled with 10 μM eFluor 670 (Cell Proliferation Dye, eBioscience, San Diego, CA, USA), and then, apoptosis was induced by UV irradiation (500 mJ/cm
2 for 2 min and 41 s) followed by 4 h of culture at 5 × 10
6/mL in IMDM in the absence of serum at 37 °C in 5% CO
2 atmosphere. Cultured PMN populations were routinely >70% early apoptotic (Annexin V
+/PI
−) and <5% late apoptotic (Annexin V
+/PI
+), as determined by flow cytometry using the Annexin V Apoptosis Detection Kit FITC (eBioscience;
Figure S5). Before the coculture with human M2(Dex) macrophages in various assays, apoptotic PMN were pretreated with either 12 μg/ml Protein S (Enzyme Research Laboratories, South Bend, IN, USA), 12.5 μg/mL MFGE8 (R&D Systems, Minneapolis, MN, USA), or 10% serum (FBS) for 1 h.
4.3. Flow Cytometry
Macrophages were detached from Nunc UpCell tissue culture plastic by incubation in cold PBS on ice for 20–30 min. After washing with ice-cold PBS containing 2% FBS, the cells (5 × 105/assay) were treated with Human BD Fc block (BD Biosciences, San Jose, CA, USA) for 10 min at room temperature followed by staining with fluorophore-conjugated monoclonal antibodies for 30 min on ice. The data from washed cells were collected using an LSRII cytometer (BD Biosciences) and analyzed with FlowJo (Treestar, Ashland, OR, USA) software. Efferocytosis receptor antibodies (R&D Systems) were PE-labeled and included the following specificities and clones: Dtk/Tyro3 (clone 96201), Axl (108724), MerTK (125518), αVβ3 (23C6), or αVβ5 (P5H9). Other antibodies and clones included PE-Cy7-conjugated antihuman CD80 (BioLegend, San Diego, CA, USA, clone 2D10), PerCp-Cy5.5-conjugated antihuman CD206 (BioLegend, 15-2), APC-Cy7-conjugated antihuman CD163 (BioLegend, GHI/61), and BV-421-conjugated antihuman CD66b (BD Biosciences, G10F5).
4.4. Efferocytosis Assay
Efferocytosis was assessed by flow cytometry [
68]. Briefly, 5 × 10
5/well monocyte-derived human M2(IL-4+IL-10+Dex) macrophages or M2(Dex) macrophages were preincubated with 10 nM recombinant PA (List Biological Laboratories, Campbell, CA, USA), 10 nM recombinant EF (List Biological Laboratories), or various concentrations of PA + EF (ET) for 4 h at 37 °C in 5% CO
2 atmosphere. Fluorescently-labeled apoptotic human PMN were preincubated with 10% serum, Protein S, or MFGE8 for 1 h unless otherwise noted and then added to the macrophages, resulting in a macrophage:PMN ratio of 1:5. After 1 h at 37 °C in 5% CO
2 atmosphere, the percentage macrophages containing intracellular (eFluor 670
+) but not surface-bound (surface CD66b
−) PMN was assessed by flow cytometry.
4.5. MerTK Inhibition Efferocytosis Assay
Human M2(Dex) macrophages (5 × 10
5/well) were preincubated with various concentrations (0–1 μM) of selective MerTK inhibitor UNC1062A [
46] for 1 h at 37 °C in 5% CO
2 atmosphere prior to the addition of fluorescently labeled, Protein S-opsonized, human apoptotic PMN. The final macrophage:PMN ratio was 1:5. After 1 h of coculture, the percentage of macrophages that had engulfed apoptotic PMN was determined by flow cytometry.
4.6. Rac Pull-Down Assay
M2(Dex) macrophages were washed four times with PBS and then incubated overnight in IMDM containing 10% FBS in the absence of M-CSF to reduce M-CSF receptor signaling to a basal state. Rested M2(Dex) macrophages were preincubated in the presence or absence of 10 nM ET for 4 h at 37 °C in 5% CO2 atmosphere, washed four times with PBS, and then cocultured (6 × 106 human M2(Dex) macrophages/well) with either Protein S- or MFGE8-opsonized human apoptotic PMN at a 5:1 PMN:M2(Dex) macrophage ratio for 15–20 min. The Rac-GTP and total Rac were measured using a Rac Activation Assay Kit (NewEast Biosciences, King of Prussia, PA, USA), following the protocol described by the manufacturer. The total lysate and Rac-GTP pulled down from lysates were electrophoresed and blotted using the manufacturer’s provided anti-Rac polyclonal antibody (2 μg/mL). The immune-reactive bands were visualized using a peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG) (1:1000 dilution) antibody (WesternSure Goat Anti-Rabbit IgG (H+L), LI-COR) and an ECL Plus Western Blotting Detection System (Azure Biosystems, Dublin, CA, USA). The images were captured using an Azure cSeries C600 (Azure Biosystems) instrument using the imaging workflow for chemiluminescent applications. The quantifications were carried out using Image J software. The ratios of Rac-GTP/Total Rac were expressed as percentages of the maximum ratio within each individual donor.
4.7. Phosphoprotein Profiling Assay
M2(Dex) macrophages were washed four times with PBS and then incubated overnight in IMDM containing 10% FBS in the absence of M-CSF to reduce M-CSF receptor signaling to a basal state. Rested M2(Dex) macrophages were preincubated in the presence or absence of 10 nM ET for 4 h at 37 °C in 5% CO2 atmosphere, washed four times with PBS, and then cocultured (6 × 106 human M2(Dex) macrophages) with Protein S-opsonized human apoptotic PMN at a 5:1 PMN:M2(Dex) macrophage ratio for 20 min. The lysates were biotinylated and used to probe the Cytoskeleton Phospho Antibody Array (Full Moon BioSystems, Sunnyvale, CA, USA) according to the manufacturer’s instructions. Biotinylated proteins captured by the arrays were detected using AlexaFluor 555-Streptavidin. The slides were scanned on a GenePix 4000B Microarray scanner at 532 nm, and the images were analyzed with GenePix Pro 7.3.1 software (Molecular Devices, San Jose, CA, USA). The fluorescence signal was obtained by calculating the Foreground (center of signal) to Background (adjacent to signal) Ratio (F/B) of each capture antibody location on the array. The ratios of signals detected by antibodies directed to phosphorylated versus un-phosphorylated proteins were compared in the presence and absence of ET.
4.8. Confocal Microscopy
M2(Dex) macrophages were washed four times with PBS, then seeded onto 8-well chambered coverglass plates (Nunc Lab-TekII Chambered Coverglass, Rochester, NY, USA, 2.5 × 105 macrophages/well), and grown overnight in IMDM containing 10% FBS at 37 °C in 5% CO2 atmosphere. The cells were preincubated with or without 10 nM ET for 4 h and then cocultured with 5-chloromethylfluorescein diacetate (CFMDA, Molecular Probes, Life Technologies, Eugene, OR, USA)-labeled apoptotic PMN that had been opsonized with Protein S at a 5:1 PMN:M2(Dex) macrophage ratio. After 30 min of co-incubation, the cells were washed three times and blocked with 10% human AB serum on ice for 15 min. CD11b was detected by incubating with AlexaFluor647-labeled antihuman CD11b antibody (Clone ICRF44, Southern Biotech, Birmingham, AL, USA) on ice for 1 h. To visualize the actin filaments, the cells were fixed with 2% para-formaldehyde and then co-stained with Phalloidin-AlexaFluor568 conjugate (Invitrogen, Eugene, OR, USA) at room temperature for 20 min and mounted in Prolong Gold with 4′,6-diamidino-2-phenylindole (DAPI) (Life Technologies). The images were captured by optical sectioning on a Zeiss LSM-510 META laser-scanning microscope and cumulatively displayed as maximal intensity projections.
4.9. Statistical Analyses
The results are presented as mean ± SEM, with the number of independent experiments or donors indicated in the figure legends. The results were analyzed by one-way ANOVA followed by a multiple comparisons posttest (p < 0.05) or Student’s t-test (p < 0.05) as indicated in the figure legends (GraphPad, San Diego, CA, USA).