Polymyxin-Induced Cell Death of Human Macrophage-Like THP-1 and Neutrophil-Like HL-60 Cells Associated with the Activation of Apoptotic Pathways

Innate immunity is crucial for the host to defend against infections, and understanding the effect of polymyxins on innate immunity is important for optimizing their clinical use. In this study, we investigated the potential toxicity of polymyxins on human macrophage-like THP-1 and neutrophil-like HL-60 cells. Differentiated THP-1 human macrophages (THP-1-dMs) and HL-60 human neutrophils (HL-60-dNs) were employed. Flow cytometry was used to measure the concentration-dependent effects (100 to 2,500 μM for THP-1-dMs and 5 to 2,500 μM for HL-60-dNs) and time-dependent effects (1,000 μM for THP-1-dMs and 300 μM for HL-60-dNs) of polymyxin B over 24 h.

ABSTRACT

Innate immunity is crucial for the host to defend against infections, and understanding the effect of polymyxins on innate immunity is important for optimizing their clinical use. In this study, we investigated the potential toxicity of polymyxins on human macrophage-like THP-1 and neutrophil-like HL-60 cells. Differentiated THP-1 human macrophages (THP-1-dMs) and HL-60 human neutrophils (HL-60-dNs) were employed. Flow cytometry was used to measure the concentration-dependent effects (100 to 2,500 μM for THP-1-dMs and 5 to 2,500 μM for HL-60-dNs) and time-dependent effects (1,000 μM for THP-1-dMs and 300 μM for HL-60-dNs) of polymyxin B over 24 h. Effects of polymyxin B on mitochondrial activity, activation of caspase-3, caspase-8, and caspase-9, and Fas ligand (FasL) expression in both cell lines were examined using fluorescence imaging, colorimetric, and fluorometric assays. In both cell lines, polymyxin B induced concentration- and time-dependent loss of viability at 24 h with 50% effective concentration (EC₅₀) values of 751.8 μM (95% confidence interval [CI], 692.1 to 816.6 μM; Hill slope, 3.09 to 5.64) for THP-1-dM cells and 175.4 μM (95% CI, 154.8 to 198.7 μM; Hill slope, 1.42 to 2.21) for HL-60-dN cells. A concentration-dependent loss of mitochondrial membrane potential and generation of mitochondrial superoxide was also observed. Polymyxin B-induced apoptosis was associated with concentration-dependent activation of all three tested caspases. The death receptor apoptotic pathway activation was demonstrated by a concentration-dependent increase of FasL expression. For the first time, our results reveal that polymyxin B induced concentration- and time-dependent cell death in human macrophage-like THP-1 and neutrophil-like HL-60 cells associated with mitochondrial and death receptor apoptotic pathways.

TEXT

In 2017, the World Health Organization (WHO) highlighted the urgent need for new antibiotics against multidrug-resistant (MDR) Gram-negative pathogens (1). Without novel classes of antibiotics in the clinic, the polymyxins (i.e., polymyxin B and colistin) are now used as a last resort against life-threatening infections caused by Gram-negative “superbugs” (2–4). Despite their excellent antibacterial activity, optimal dosing of intravenous polymyxins is limited by their propensity to cause nephrotoxicity, which may occur in up to 60% of patients (5–7). Unlike intravenous administration, inhaled polymyxins can achieve much higher exposure in the lungs for the treatment of pulmonary infections, with minimal systemic exposure and unwanted nephrotoxicity (8–12). For example, concentrations of formed colistin in the tracheal aspirate of neonates experiencing ventilator-associated pneumonia (mean ± standard deviation [SD], 24 ± 8.2 μg/ml) were much higher than those in plasma (0.59 ± 0.35 μg/ml) following a single dose (120,000 IU/kg, equivalent to 9.6 mg/kg) of nebulized colistimethate sodium (8). Similarly, concentrations of formed colistin in critically ill patients treated with inhaled colistimethate sodium (daily dose ranging from 240 to 400 mg [i.e., 3 to 5 million IU]) were much higher (up to 1,550-fold) in the lung epithelial lining fluid (ELF) than in plasma (9, 10, 13). Notwithstanding the targeting advantage of direct administration to the lungs, inhalation of high-dose polymyxins may potentially affect the innate immune system within the lungs (14). As innate immunity is a major component of the host defense against microbial pathogens (15, 16), the impact of polymyxins on innate immune cells (e.g., macrophages and neutrophils) should be examined in order to optimize their inhalation therapy.

Following bacterial infection, innate immune cells (e.g., macrophages and neutrophils) are recruited to the infection site to eradicate the pathogens (17–20). Several antibiotics (e.g., macrolides and carbapenems) have been reported to impact the immune response through modulating cellular respiration of immune cells and their antibacterial killing activity and altering their chemotaxic responses, thereby affecting disease progression (21–24). The relationship between the effect of antibiotics (e.g., macrolides, tetracyclines, aminoglycosides, and rifampin) on the immune response and bacterial resistance has been investigated (25), and it is purported that a synergistic effect of antimicrobial therapy and the immune system is crucial for favorable treatment outcomes (24, 26, 27).

Apoptosis plays an essential role in regulating immunological responses of macrophages and neutrophils (28, 29). Interruption of programmed cell death of macrophages and neutrophils is connected with a disturbance of their function (30). Notably, we have previously shown that polymyxins can induce apoptosis in kidney proximal tubular and lung epithelial cells (31, 32). To date, the impact of polymyxin-induced immunotoxicity on human macrophages and neutrophils has not been investigated. Our present study demonstrated that polymyxin B induced a concentration- and time-dependent cell death in macrophage-like THP-1 and neutrophil-like HL-60 cells via both intrinsic and extrinsic apoptotic pathways.

RESULTS

Concentration- and time-dependent cell death in macrophage-like THP-1-dM and neutrophil-like HL-60-dN cells induced by polymyxin B treatment.

In the first series of experiments, we investigated the effect of polymyxin B concentration on macrophage-like THP-1-dM and neutrophil-like HL-60-dN cell viability. The polymyxin B concentrations were clinically relevant and correspond to the concentrations achieved in the lung epithelial lining fluid of patients treated with inhaled polymyxins (10). Neutrophil-like HL-60-dN cells were more susceptible to polymyxin B than macrophage-like THP-1-dM cells under the same treatment conditions. Polymyxin B induced a concentration-dependent cell death in both macrophage-like THP-1-dM and neutrophil-like HL-60-dN cells (Fig. 1A and B). Following 24 h of treatment with polymyxin B at 200 μM and 1,000 μM, cell death (%) of THP-1-dMs reached 14.3% ± 2.88% and 67.6% ± 12.8%, respectively. For HL-60-dN cells, treatment with 50 μM and 200 μM polymyxin B for 24 h resulted in 20.6% ± 3.97% and 58.7% ± 4.42% cell death, respectively. The 50% effective concentration (EC₅₀) values for polymyxin B treatment were 751.8 μM (95% CI, 692.1 to 816.6 μM; Hill slope, 3.09 to 5.64) for THP-1-dM cells and 175.4 μM (95% CI, 154.8 to 198.7 μM; Hill slope, 1.42 to 2.21) for HL-60-dN cells (Fig. 1B). Following treatment with 100 μM polymyxin B for 24 h, the viability of HL-60-dN cells decreased to approximately 30% while the viability of THP-1-dM remained similar to that in the control group (Fig. 1C). In the time-dependent studies (Fig. 2), the percentage of THP-1-dM cell deaths at 1 h was 17.3% ± 4.55%, which increased to 66.0% ± 11.4% at 24 h following treatment with 1,000 μM polymyxin B; the corresponding values for HL-60-dN cells following treatment with 300 μM polymyxin B were 6.15% ± 1.19% and 75.4% ± 12.1%.

FIG 1.

Polymyxin B-induced cell death in macrophage-like THP-1 and neutrophil-like HL-60 cells. (A) Propidium iodide (PI)-positive cells of polymyxin B-treated macrophage-like THP-1 and neutrophil-like HL-60 cells determined by flow cytometry at 24 h. (B) Cell death (%) of macrophage-like THP-1 (filled circles) and neutrophil-like HL-60 cells (filled triangles) as a function of polymyxin B concentration at 24 h. (C) Viability of macrophage-like THP-1 (filled circles) and neutrophil-like HL-60 (filled triangles) cells in the presence of polymyxin B at 24 h. Data are presented as mean ± standard deviation (SD) (n = 3). ****, P < 0.0001.

FIG 2.

Polymyxin B-induced time-dependent loss of viability in macrophage-like THP-1 and neutrophil-like HL-60 cells. (A) Propidium iodide (PI)-positive cells of polymyxin B-treated macrophage-like THP-1 and neutrophil-like HL-60 cells determined at different time points (1, 4, 12, and 24 h) using flow cytometry. (B) Time-dependent cell death of macrophage-like THP-1 (filled circles) and neutrophil-like HL-60 (filled triangles) cells incubated with polymyxin B (1,000 and 300 μM, respectively). Control values of macrophage-like THP-1 and neutrophil-like HL-60 cells are presented with nonfilled circles and nonfilled triangles, respectively. Data are presented as mean ± SD (n = 3).

Mitochondrial membrane potential, superoxide formation, and oxidoreductase activity induced by polymyxin B treatment in macrophage-like THP-1-dM and neutrophil-like HL-60-dN cells.

Polymyxin B treatment led to a concentration-dependent loss of mitochondrial membrane potential in both macrophage-like THP-1-dM and neutrophil-like HL-60-dN cells (Fig. 3). Compared to the untreated control cells, the intensity of tetramethylrhodamine, ethyl ester (TMRE) fluorescence decreased to 38.4% ± 11.2% and 14.3% ± 7.01% for the THP-1-dM cells, and 46.3% ± 14.1% and 14.0% ± 7.13% for the HL-60-dN cells, following 24 h of treatment with 500 μM and 1,000 μM polymyxin B, respectively (Fig. 3B). Similarly, a concentration-dependent formation of mitochondrial superoxide was observed in both macrophage-like THP-1-dM and neutrophil-like HL-60-dN cells following polymyxin B treatment (Fig. 4). In THP-1-dM cells, polymyxin B at 200 μM and 500 μM induced 4.1- and 8.8-fold increases, respectively, in the fluorescence intensity (MitoSOX) compared to that of untreated control cells; the corresponding values in HL-60-dN cells at the same polymyxin B concentrations were 1.7- and 4.5-fold. A concentration-dependent decrease in oxidoreductase activity was also observed in both cell lines (Fig. 5).

FIG 3.

Polymyxin B damages mitochondria in macrophage-like THP-1 and neutrophil-like HL-60 cells. (A) Concentration-dependent loss of the mitochondrial membrane potential in tetramethylrhodamine, ethyl ester (TMRE) and Hoechst 33342-stained macrophage-like THP-1 and neutrophil-like HL-60 cells following 24 h of treatment with polymyxin B. (B) Quantification of TMRE-positive macrophage-like THP-1 (filled circles) and neutrophil-like HL60 (filled triangles) cells following polymyxin B treatment. Data are presented as mean ± SD (n = 3). Bar, 50 μm.

FIG 4.

Polymyxin B induces mitochondrial superoxide formation in macrophage-like THP-1 and neutrophil-like HL-60 cells. (A) Concentration-dependent formation of mitochondrial superoxide following 24 h of treatment with polymyxin B in macrophage-like THP-1 and neutrophil-like HL-60 cells. (B) Quantification of MitoSOX fluorescence intensity in polymyxin B-treated macrophage-like THP-1 (filled circles) and neutrophil-like HL-60 (filled triangles) cells. Data are presented as mean ± SD (n = 3). Bar, 50 μm.

FIG 5.

Mitochondrial toxicity (XTT [2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide salt] assay) in macrophage-like THP-1 (filled circles) and neutrophil-like HL-60 (filled triangles) cells following 24 h of treatment with polymyxin B. Data are presented as mean ± SD (n = 3).

Polymyxin B-activated caspase-3, caspase-8, and caspase-9 and induced FasL expression in macrophage-like THP-1-dM and neutrophil-like HL-60-dN cells.

Polymyxin B treatment led to a concentration-dependent activation of caspase-3, caspase-8, and caspase-9 (Fig. 6). Compared to the untreated controls, macrophage-like THP-1-dM cells treated with 1,000 μM polymyxin B showed an ∼10.5-fold increase in caspase-3 activity (Fig. 6A). Caspase-8 and caspase-9 activities also increased by approximately 9.5- and 5.5-fold, respectively, following equivalent treatment. In neutrophil-like HL-60-dN cells treated with 100 μM polymyxin B (Fig. 6B), the activity of caspase-3, caspase-8, and caspase-9 increased by approximately 25.4-, 54.9-, and 14.2-fold, respectively. Similarly, FasL was expressed in a concentration-dependent manner in both macrophage-like THP-1-dM (Fig. 7A) and neutrophil-like HL-60-dN cells (Fig. 7B) with polymyxin B treatment.

FIG 6.

Activation of caspase-3 (filled circles), caspase-8 (filled squares), and caspase-9 (filled triangles) in (A) macrophage-like THP-1 and (B) neutrophil-like HL-60 cells following 18 h of treatment with polymyxin B. Data are presented as mean ± SD (n = 3).

FIG 7.

(A) FasL expression in macrophage-like THP-1 and neutrophil-like HL-60 cells following 24 h of treatment with different concentrations of polymyxin B. (B) FasL expression in polymyxin B-treated macrophage-like THP-1 (filled circles) and neutrophil-like HL-60 cells (filled triangles) was compared to that of the control groups (presented with nonfilled circles in macrophage-like THP-1 cells and nonfilled triangles in neutrophil-like HL-60 cells). Data are presented as mean ± SD (n = 3).

DISCUSSION

In light of the important role of pulmonary innate immunity in host protection from bacterial infections, an understanding of the potential immune toxicity of aerosolized antibiotics such as polymyxins is essential (14). Apoptosis is a form of programmed cell death that has a key role in immune regulation, as controlling the death of immune cells is essential for preventing the destruction of healthy tissues (30). Several antibiotics are able to modulate apoptosis signals in macrophages and neutrophils, which consequently affect both the host and pathogen (33–36). The regulatory role of apoptosis is significant for the immune response, and disturbances of apoptosis in macrophages and neutrophils have been linked with serious clinical conditions (28, 37–39). THP-1 and HL-60 cell lines have been extensively employed for research on human macrophages and neutrophils, respectively (40, 41). In the present study, we examined the impact of polymyxin B on human macrophage-like THP-1 and neutrophil-like HL-60 cells at concentrations achieved in the lung epithelial lining fluid in patients (up to 1,137 mg/liter) following inhalation administration (10). At our chosen (clinically relevant) concentrations, polymyxin B induced cell death in both cell types in a concentration- and time-dependent manner (Fig. 1 and 2). Given that the only other polymyxin in clinical use, colistin (polymyxin E), differs from polymyxin B by only a single amino acid (42–44), it is very likely that the effects of colistin on immune cells would be similar to those reported here for polymyxin B. Our results indicate that the potential toxic effect of polymyxins on pulmonary macrophages and neutrophils needs to be considered when optimizing the dosage regimens of inhaled polymyxins.

Interestingly, the EC₅₀ values of polymyxin B in macrophage-like THP-1-dM (751.8 μM) and neutrophil-like HL-60-dN (175.4 μM) cells were substantially different (Fig. 1B). These different sensitivities can be explained in light of the disparate responses of the two different types of immune cells to inflammatory stimuli (29, 45, 46). Neutrophils act in a suicidal manner and undergo self-triggered apoptosis with a short turnover rate (6 to 12 h) that becomes even shorter following treatment with apoptotic agents (29). In contrast, macrophages are more robust, with a much longer life span and different survival mechanisms to resist apoptotic stimuli (45, 46). A similar pattern of differential cell killing has recently been reported, in which the macrolide antibiotic tylvalosin-induced rapid apoptosis in treated neutrophils (0.5 and 1 h posttreatment) but not macrophages (12 and 24 h posttreatment) (33). Our data show that neutrophil-like HL-60-dN are much more sensitive to polymyxin B than macrophage-like THP-1-dM cells (Fig. 1C).

A key early event that triggers apoptosis in mammalian cells is mitochondrial outer membrane permeabilization (47). Apoptotic death of neutrophils and macrophages involves loss of mitochondrial outer membrane integrity, release of mitochondrial intermembrane space proteins, and caspase activation (28, 29). As neutrophils mainly rely on glycolysis for energy generation (48), we suggested that polymyxin B-associated mitochondrial dysfunction is primarily linked to irregular mitochondrial respiration and apoptosis initiation. Polymyxin B induced a concentration-dependent loss of mitochondrial membrane potential (ΔΨ) (Fig. 3), generation of reactive oxygen species (ROS) (Fig. 4), and decreased activity of cellular respiration oxidoreductases (Fig. 5). Of note, polymyxins are able to induce mitochondrial stress in human kidney proximal tubular (HK-2) and lung epithelial (A549) cells (32, 49). Collectively, our results demonstrate that mitochondrial toxicity plays an important role in polymyxin-induced cell death in macrophage-like THP-1-dM and neutrophil-like HL-60-dN cells. Systems pharmacological studies are required to further elucidate the mechanisms underpinning mitochondrial toxicity and are under way in our laboratory.

Two main apoptotic pathways, extrinsic (death receptor) and intrinsic (mitochondrial), are present in mammalian cells, including in macrophages and neutrophils (50–52). The extrinsic pathway is activated through the ligation of surface death receptors with FasL, tumor necrosis factor alpha (TNF-α), or TNF-related apoptosis-inducing ligand (TRAIL), and the formation of death receptor-ligand complexes that recruit and activate caspase-8 (53). The intrinsic pathway is regulated at the mitochondrial level; mitochondrial membrane permeabilization initiates the release of proapoptotic factors from the mitochondrial interspace that eventually lead to the recruitment and activation of caspase-9 (54). In the present study, polymyxin treatment substantially increased the activation of caspase-3, caspase-8, and caspase-9 in both macrophage-like THP-1-dM and neutrophil-like HL-60-dN cells (Fig. 6). Compared to their control samples, the fold increase in the activity of each caspase in HL-60-dNs in the presence of 100 μM polymyxin B was higher than that in THP-1-dMs (Fig. 6), which supports their differential susceptibility to polymyxin B treatment, as indicated by their corresponding EC₅₀ values (Fig. 1B). The concentration-dependent stimulation of all three caspases reveals that both mitochondria and death receptor-mediated apoptotic pathways were involved in the polymyxin B-induced cell death in human macrophage-like THP-1 and neutrophil-like HL-60 cells. In both types of innate immune cells, a greater activation of caspase-8 than of caspase-9 was observed under the experimental conditions in our study, suggesting a primary involvement of the extrinsic pathway in polymyxin-induced apoptosis. The Fas death receptor/Fas ligand (FasL) system is an essential regulator of cell death in macrophages and neutrophils (50, 51). In the present study, a concentration-dependent increase of FasL expression was evident in both macrophage-like THP-1 and neutrophil-like HL-60 cells (Fig. 7). Ligation of FasL with Fas induces a conformational change in the plasma membrane that allows recruitment of FAS-associated protein with death domain (FADD), which is the principal adaptor protein responsible for transmitting death receptor-mediated apoptotic signals (55). Together with the activation of caspase-8, our results indicate that the death receptor apoptosis pathway plays a major role in polymyxin-induced apoptosis in macrophage-like THP-1 and neutrophil-like HL-60 cells. According to the apoptosis-based cellular classification, neutrophils are classified as type II cells in which the Fas-triggered apoptosis requires an amplification loop usually mediated by the caspase-9-activated mitochondrial apoptotic pathway (52). The induction of death receptors mediates apoptosis, and the resultant activation of caspase-8 accelerates the release of mitochondrial apoptotic mediators (e.g., cytochrome c) which in turn triggers the activation of caspase-9 (56). Our observations are similar to those with human kidney proximal tubular cells, in which the activity of caspase-8 was greater than caspase-9 (49), but different from the pattern of caspase activation reported in polymyxin-treated human lung epithelial A549 cells (32). This may indicate different cellular responses to polymyxin treatment in different cells. The activation of both caspase-8 and caspase-9 in our study (Fig. 6) suggests an interplay between the extrinsic and intrinsic pathways of polymyxin-induced apoptosis in human macrophage-like THP-1 and neutrophil-like HL-60 cells. The kinetics of the activated apoptotic mediators warrant further investigations in order to elucidate the mechanism of polymyxin-induced apoptosis in innate immune cells.

To the best of our knowledge, this is the first study to demonstrate that polymyxin B produces a concentration- and time-dependent apoptotic effect on human macrophage-like THP-1 and neutrophil-like HL-60 cells. These findings highlight that both mitochondrial and death receptor apoptotic pathways are involved in the toxicity of polymyxins on cells of the innate immune system. Our study provides important pharmacological information for optimizing polymyxin inhalation therapy in patients.

MATERIALS AND METHODS

Chemicals and reagents.

A stock solution of 20.0 mM polymyxin B sulfate (catalog number 86-40302, 98% purity, ≥95% potency; Beta Pharma, Shanghai, China) was prepared in Milli-Q water and sterilized by a syringe filter (0.22 μM, Millex-GV; Millipore). A stock solution (34.0 mM) of etoposide (catalog number E1383, ≥98% potency; Sigma-Aldrich, NSW, Australia) was prepared in dimethyl sulfoxide (DMSO; Sigma-Aldrich).

Cell culture.

Human acute monocytic leukemia cell line (THP-1) and human leukemia cell line (HL-60) from the American Type Culture Collection (ATCC) (passages 3 to 15; Manassas, VA, USA) were employed as models for macrophages and neutrophils, respectively. Both cell lines were grown in RPMI 1640 medium supplemented with 25 mM HEPES (Media Preparation Unit, Biomedicine Discovery Institute, Monash University, Clayton, Australia) and 10% fetal bovine serum (FBS) (Invitrogen, Life Technologies, Victoria, Australia).

Cell differentiation.

Differentiated macrophage-like THP-1 and neutrophil-like HL-60 cells are widely used to represent the morphological and functional features of human macrophages and neutrophils, respectively (40, 41). In the present study, THP-1 and HL-60 cells were differentiated as reported previously (41). Briefly, both types of cells (2 × 10⁵ cells/ml) were seeded into 24-well (1-ml) and 96-well (0.2-ml) plates at 37°C in a humidified atmosphere containing 5% CO₂. THP-1 cells were induced to differentiate into macrophage-like cells (THP-1-dMs) after treatment with 25 nM phorbol 12-myristate 13-acetate (PMA) (catalog number sc-3576A; Thermo Fisher Scientific, Victoria, Australia) for 48 h, followed by medium removal and addition of fresh medium for another 24 h to allow cellular recovery (41, 57). HL-60 cells were treated with 1.25% dimethyl sulfoxide (DMSO) for 72 h to differentiate into neutrophil-like cells (HL-60-dNs) (58). Polymyxin B was added into RPMI 1640 supplemented with 10% FBS to treat differentiated cells.

Assessments of concentration- and time-dependent cell death due to polymyxin B.

THP-1 and HL-60 cells cultured and differentiated in 24-well plates were incubated for 24 h in the presence and absence of polymyxin B (100 to 2,500 μM for THP-1-dMs and 5 to 2,500 μM for HL-60-dNs) to evaluate concentration-dependent cell death (31). For THP-1-dMs, incubated cells were detached using TrypLE Select (10×) (Gibco, Life Technologies Australia Pty Ltd.). For HL-60-dN cells, no detachment was required, as the differentiated cells were nonadherent. Cells were stained with propidium iodide (PI) (1.5 μM, red fluorescence, excitation [Ex]/emission [Em] = 493/535 to 617 nm; Invitrogen), and viability was measured using flow cytometry (Novocyte flow cytometer; ACEA Biosciences, Inc., San Diego, CA) (31). The concentration required to induce 50% cell death by polymyxin B (EC₅₀) was calculated by fitting data to a Hill slope function using unweighted nonlinear least squares regression analysis (Prism v7.0; GraphPad Software, San Diego, CA) (59). Time-dependent cell death was measured in the presence of 1,000 μM polymyxin B for THP-1-dM cells and 300 μM for HL-60-dN cells at 1, 4, 8, 12, 18, and 24 h using fluorescence-activated cell sorting (FACS) as described above.

Loss of mitochondrial membrane potential and formation of mitochondrial superoxide.

Differentiated macrophage-like THP-1 and neutrophil-like HL-60 cells in 96-well plates were treated for 24 h with polymyxin B (100 to 1,000 μM for THP-1-dMs and 50 to 1,000 μM for HL-60-dNs); the control was treated with Milli-Q water only. Loss of mitochondrial membrane potential (ΔΨ) and formation of mitochondrial superoxide were examined using tetramethylrhodamine ethyl ester stain (TMRE) (50 nM; Ex/Em = 561/568 to 690 nm; Invitrogen) and MitoSOX red dye (5 μM, Ex/Em = 514/531 to 622 nm; Invitrogen), respectively (49). The fluorescence intensity was quantified using an EVOS FL Auto imaging system (Life Technologies Australia Pty Ltd.) (60). Total cell counting was conducted using 4′,6-diamidino-2-phenylindole (Hoechst 33342, 2 μg/ml, Ex/Em = 350/390 to 551 nm; Invitrogen) (49). The average fluorescence intensity per cell for each treatment was calculated using ImageJ and presented as a percentage of fluorescence intensity of the control group (59, 61). An average of at least 75 cells/field and three different fields per condition were analyzed for each condition.

Measurement of mitochondrial toxicity.

Differentiated macrophage-like THP-1 and neutrophil-like HL-60 cells in 96-well plates were treated for 24 h with polymyxin B (100 to 1,250 μM for THP-1-dMs and 50 to 1,000 μM for HL-60-dNs); the control was treated with Milli-Q water. Cell-free wells containing culture medium only were prepared for measuring blank absorbance readings. Following the polymyxin B treatment, tetrazolium dye (XTT [2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide salt] sodium salt; Santa Cruz Biotechnology, Dallas, USA) combined with phenazine methosulfate (PMS, ≥98% purity, as an intermediate electron receptor; Sigma-Aldrich) was employed to measure the oxidoreductase activity (62). Cells from each group were prepared and treated in triplicates. The plate was wrapped with aluminum foil and incubated at 37°C for 4 h with gentle shaking every 30 min to evenly distribute the color. Absorbance was measured at 475 and 660 nm using an Infinite M200 plate reader (Tecan Group, Ltd., Zürich, Switzerland) and the absorbance difference (A₄₇₅ − A₆₆₀) for each group was calculated. Percentages of the measured absorbance difference (A₄₇₅ − A₆₆₀) from control values were plotted against the corresponding polymyxin B concentration.

Assessment of caspase-3, caspase-8, and caspase-9 activity.

A multiplex activity assay kit (Abcam, Victoria, Australia) consisting of fluorogenic cell membrane-permeable caspase-specific substrates was employed to assess the activation of caspase-3, caspase-8, and caspase-9 due to polymyxin B treatments (63). Briefly, THP-1 and HL-60 cells were seeded and differentiated in 96-well black plates (90 μl/well and 2.5 × 10⁴ cells/well) and treated for 18 h with polymyxin B (100 to 1,000 μM for THP-1-dMs and 25 to 200 μM for HL-60-dNs); the control was treated with Milli-Q water. Cells incubated with etoposide (final concentration of 340 μM) were used as the positive control. Culture medium was used for measuring blank absorbance readings. Substrates of caspase-3, caspase-8, and caspase-9 were all diluted and mixed together in the provided assay buffer (1:200) before adding 100 μl of the caspase assay loading solution to each well. The plate was protected from light and incubated at 37°C for 45 min. Fluorescence was measured using the microplate reader as above at the following wavelengths: caspase-3, Ex/Em = 535/620 nm; caspase-8: Ex/Em = 490/525 nm; caspase-9: Ex/Em = 370/450 nm. Blank readings were subtracted from the values of treated and control groups, and fold change in the fluorescence intensity was measured between different groups.

Measurement of FasL expression.

Expression of polymyxin B-induced Fas ligand (FasL) in macrophage-like THP-1-dM and neutrophil-like HL-60-dN cells was determined using flow cytometry. Differentiated cells in 24-well plates were treated for 24 h with polymyxin B (100 to 2,000 μM for THP-1-dMs and 25 to 1,000 μM for HL-60-dNs); the control group was treated with Milli-Q water only. After treatment, cells were detached and collected as described above. Cell suspension was centrifuged, and cell pellets were resuspended with phosphate-buffered saline (PBS, pH 7.4; Media Preparation Unit). Subsequently, cells were incubated with Alexa Fluor 647-conjugated anti-CD178 (FasL) monoclonal antibody (MAb) (mouse anti-human CD178, Alexa Fluor 647; AbD Serotec, Germany) for 45 min at room temperature. After incubation, cells were centrifuged (22°C, 140 × g, 5 min) and resuspended with PBS at room temperature. FasL expression in cells was measured by a FACS cytometer as above (Ex/Em = 650/665 nm).

Statistical analysis.

All experiments were conducted with three independent replicates and results are presented as mean ± standard deviation (SD). One-way analysis of variance (ANOVA) and Tukey’s test with a significance level of P < 0.05 were performed using GraphPad Prism (v7.0).