Potential Toxicity of Polymyxins in Human Lung Epithelial Cells

Abstract

Inhaled polymyxins are of considerable utility in achieving optimal exposure in the respiratory tract for the treatment of lung infections caused by multidrug-resistant Gram-negative pathogens. Current inhaled polymyxin therapy is empirical, and often large doses are used that may lead to potential pulmonary adverse effects. This study aimed to investigate the effect of polymyxins on human lung epithelial (A549) cells. The viability of A549 cells was examined after treatment with polymyxins by flow cytometry. Activation of caspases 3, 8, and 9, expression of Fas ligand (FasL), loss of mitochondrial membrane potential, and mitochondrial oxidative stress induced by polymyxin B were evaluated. The concentration of polymyxin B required to induce 50% of maximal cell death was 1.74 mM (95% confidence interval, 1.60 to 1.90 mM). Colistin was at least 2-fold less toxic than polymyxin B, while colistimethate was nontoxic. With 2.0 mM polymyxin B, 30.6% ± 11.5% (mean ± standard deviation) of the cells were apoptotic at 8 h and this increased to 71.3% ± 3.72% at 24 h. Concentration- and time-dependent activation of caspases 3, 8, and 9 was evident, while the activation of caspase 9 was more dramatic. Furthermore, polymyxin B caused concentration- and time-dependent FasL expression, production of mitochondrial reactive oxygen species, and changes in mitochondrial membrane potential. This is the first study to demonstrate that both extrinsic death receptor and intrinsic mitochondrial pathways are involved in polymyxin-induced toxicity in A549 cells. This knowledge base is critical for the development of novel strategies for the safe and effective inhalation therapy of polymyxins against Gram-negative "superbugs."

Keywords: apoptosis; mitochondria; polymyxin; pulmonary delivery; respiratory toxicity.

FIG 1

A549 cell death and apoptosis induced by polymyxins. (A) A549 cell death at 24 h as a function of polymyxin B concentration. Data are presented as the mean ± SD (n = 3). The EC₅₀ values of colistin and CMS were not calculated, as colistin-induced cell death (red) did not reach a plateau and no significant cell death was observed with CMS treatment (green). (B) Time-dependent cell death induced by polymyxin B at 2.0 mM (mean ± SD; n = 3). Filled circles represent polymyxin B treatment, and filled squares represent the untreated control.

FIG 2

(A) Sensitivity of A549 (black bars) and HK-2 (gray bars) cells to polymyxin B. ****, P < 0.0001. (B) Polymyxin B distribution in HK-2 and A549 cells treated with 12.5 μM polymyxin B for 24 h with an anti-polymyxin B MAb. Scale bar, 10 μm.

FIG 3

Concentration (A, B)- and time (C, D)-dependent activation of caspase 3 in A549 cells. Activation was measured with the caspase 3-specific fluorogenic substrate Red-DEVD-FMK. For the time-dependent experiments, the black and gray bars represent 1.0 and 2.0 mM polymyxin B, respectively. Scale bars, 50 μm. Group results are presented as the mean ± SD; n = 3. ***, P < 0.001; ****, P < 0.0001 compared with control samples. In panel B, the concentration-dependent data represent the 24-h time point.

FIG 4

Concentration (A, B)- and time (C, D)-dependent activation of caspase 8 in A549 cells measured with the caspase 8-specific fluorogenic substrate Red-IETD-FMK. In the time-dependent experiments, the black and gray bars represent 1.0 and 2.0 mM polymyxin B, respectively. Scale bars, 50 μm. Group results are presented as the mean ± SD; n = 3. ***, P < 0.001; ****, P < 0.0001 compared to control samples. In panel B, the concentration-dependent data represent the 24-h time point.

FIG 5

Concentration (A, B)- and time (C, D)-dependent activation of caspase 9 in A549 cells. Activation was measured with the caspase 9-specific fluorogenic substrate Red-LEHD-FMK. In the time-dependent experiments, the black and gray bars represent 1.0 and 2.0 mM polymyxin B, respectively. Scale bars, 50 μm. Group results are presented as the mean ± SD; n = 3. *, P < 0.05; **, P < 0.01; ****, P < 0.0001 compared with control samples. In panel B, the concentration-dependent data represent the 24-h time point.

FIG 6

(A, B) Concentration- and time-dependent FasL expression in A549 cells treated with 0.25, 1.0, or 2.0 mM polymyxin B for 24 h. FasL expression was detected by flow cytometry with anti-FasL antibody CD178 conjugated with Alexa Fluor 647. (C, D) Time-dependent FasL expression of A549 cells treated with 1.0 mM (black bars) or 2.0 mM (gray bars) polymyxin B. The profiles are representative of three independent experiments. The percentage of FasL-positive cells is presented as the mean ± SD (n = 3). *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 compared with control samples. In panel B, the concentration-dependent data represent the 24-h time point.

FIG 7

Polymyxin B-induced mitochondrial oxidative stress in A549 cells. (A, B) Concentration-dependent (0.25, 1.0, and 2.0 mM) polymyxin-induced mitochondrial oxidative stress was detected with MitoSOX. (C, D) Time-dependent mitochondrial oxidative stress at 1.0 mM (black bars) or 2.0 mM (gray bars) polymyxin B. Scale bars, 50 μm. Results are presented as the mean ± SD (n = 3). **, P < 0.01; ****, P < 0.0001 compared with control samples. In panel B, the concentration-dependent data represent the 24-h time point.

FIG 8

Colistin- and CMS-induced mitochondrial oxidative stress in A549 cells. (A, B) Concentration-dependent (2.0, 4.0, and 6.0 mM) mitochondrial oxidative stress was detected with MitoSOX only in the colistin group at 24 h. Scale bar, 50 μm. Results are presented as the mean ± SD (n = 3). **, P < 0.01; ****, P < 0.0001 compared with control samples.

FIG 9

Polymyxin B-induced loss of mitochondrial membrane potential in A549 cells. (A, B) Concentration-dependent (0.25, 1.0, and 2.0 mM) polymyxin-induced loss of mitochondrial membrane potential detected with TMRE. (C, D) Time-dependent loss of mitochondrial membrane potential in A549 cells treated with 1.0 mM (black bars) or 2.0 mM (gray bars) polymyxin B for 24 h. Scale bars, 50 μm. Results are presented as the mean ± SD (n = 3). ****, P < 0.0001 compared with control samples. In panel B, the concentration-dependent data represent the 24-h time point.