Effect of Sub-MICs of Macrolides on the Sensitivity of Pseudomonas aeruginosa to Nitrosative Stress: Effectiveness against P. aeruginosa with and without Multidrug Resistance

Sub-MICs of the 14-membered macrolides erythromycin (EM) and clarithromycin (CAM) decreased the growth of Pseudomonas aeruginosa PAO1 and increased its sensitivity to endogenous and exogenous nitrosative stress. However, a 16-membered macrolide, josamycin (JM), was not or less effective. In 9 of 13 non-multidrug-resistant P. aeruginosa (non-MDRP) and 9 of 27 MDRP ST235 strains, the sub-MIC of EM induced significant reductions in bacterial numbers following treatment with a nitric oxide donor.

ABSTRACT

Sub-MICs of the 14-membered macrolides erythromycin (EM) and clarithromycin (CAM) decreased the growth of Pseudomonas aeruginosa PAO1 and increased its sensitivity to endogenous and exogenous nitrosative stress. However, a 16-membered macrolide, josamycin (JM), was not or less effective. In 9 of 13 non-multidrug-resistant P. aeruginosa (non-MDRP) and 9 of 27 MDRP ST235 strains, the sub-MIC of EM induced significant reductions in bacterial numbers following treatment with a nitric oxide donor.

INTRODUCTION

Pseudomonas aeruginosa is one of the most important bacterial pathogens in patients with chronic pulmonary diseases, such as diffuse panbronchiolitis (DPB) and cystic fibrosis (CF) (1). Although P. aeruginosa is generally highly resistant to macrolides, several studies have highlighted the benefits of long-term and low-dose macrolide treatment in patients with DPB or CF (2–4). Because the levels of macrolides used during low-dose treatment are too low to have sufficient antimicrobial effects, the mechanisms through which macrolides affect the outcome of chronic infections with P. aeruginosa could include modulation of the production of bacterial virulence factors and/or anti-inflammatory activities (5). To investigate the effect of sub-MICs of macrolides on the sensitivity of P. aeruginosa to nitrosative stress, we compared the sensitivities of non-multidrug-resistant P. aeruginosa (non-MDRP), MDRP, and MDRP ST235 strains against nitric oxide (NO) following macrolide treatment.

The P. aeruginosa strains used in this study were collected from RIKEN BRC (PAO1), Department of Clinical Laboratory, Chiba University Hospital (1106 and 972), and the Laboratory of Bacterial Drug Resistance, Gunma University. The MICs of erythromycin (EM), clarithromycin (CAM), and josamycin (JM) for P. aeruginosa PAO1 were 256, 256, and >256 μg/ml, respectively (6). P. aeruginosa PAO1 was grown in LB broth containing macrolides at various concentrations at 37°C. Exposure to EM or CAM for 18 h slightly decreased the bacterial numbers of P. aeruginosa PAO1 (Fig. 1A to D). However, JM did not lead to a decrease in bacterial numbers (Fig. 1E and F). Furthermore, exposure to the aminoglycoside gentamicin (GM) (MIC, 4 μg/ml) did not result in a decrease in bacterial numbers of P. aeruginosa PAO1 (Fig. 1G and H).

FIG 1.

Effect of macrolides on the growth of P. aeruginosa PAO1 and its sensitivity to endogenous and exogenous nitrosative stresses. P. aeruginosa PAO1 was grown in LB broth containing macrolides at various concentrations at 37°C for 18 h (A to H, ●). For the endogenous nitrosative stress assay (ACEG), the 18-h grown bacteria were incubated in a saline solution containing 0.1% glucose, 10 mM MOPS (pH 5.8), and macrolides at various concentrations with (■) or without (▲) 3 mM nitrite at 37°C for 22 h. For the exogenous nitrosative stress assay (BDFH), the 18-h grown bacteria were incubated in a saline solution containing 0.1% glucose, 10 mM MOPS (pH 7.0), and macrolides at various concentrations with (■) or without (▲) 100 μM DETA/NO at 37°C for 22 h. The bacterial number was determined using bacterial plate counts (CFU). All the assays were repeated three times independently. The results are expressed as means ± standard errors.

Endogenous NO is produced from nitrite as a by-product of nitrite reduction to ammonia in bacteria (7, 8). Because overexpression of nitrite reductase in P. aeruginosa is lethal (9, 10), endogenous NO is also potentially toxic in bacterial cells. Moreover, the airway mucus of CF patients is rich in nitrite and nitrate (11). In the absence of macrolides, bacterial numbers of P. aeruginosa PAO1 that were incubated in saline solution containing 0.1% glucose and 10 mM morpholinepropanesulfonic acid (MOPS; pH 5.8) and treated with 3 mM nitrite were in the range of 2 × 10⁵ to 4 × 10⁵ CFU/100 μl after 22 h (Fig. 1A, C, E, and G). In the presence of macrolides, bacterial numbers of P. aeruginosa PAO1 treated with 3 mM nitrite were decreased by EM and CAM in a dose-dependent manner (Fig. 1A and C), whereas there were no changes in bacterial numbers of P. aeruginosa PAO1 treated with both 3 mM nitrite and JM (Fig. 1E). Bacterial numbers of P. aeruginosa PAO1 treated with 3 mM nitrite also did not decrease by exposure to GM (Fig. 1G).

Exogenous NO is produced by host defense mechanisms (12), and it is synthesized at high concentrations by inducible NO synthases in phagocytic cells (13–16). Furthermore, because NO is a water-soluble free radical, it can diffuse into bacterial cells from outside the bacteria and pass through the biofilm. The concentrations of NO generated from 100 μM NO donor (Z)-1-[N-(2-aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium-1,2-diolate (DETA/NO; Cayman Chemical Company, Ann Arbor, MI) corresponded to NO concentrations detected under physiological conditions (17). In the absence of macrolides, the bacterial numbers of P. aeruginosa PAO1 that were incubated in saline solution containing 0.1% glucose and 10 mM MOPS (pH 7.0) and treated with an NO donor were 1 × 10⁴ to 4 × 10⁵ CFU/100 μl after 22 h (Fig. 1B, D, F and H). Unexpectedly, in the presence of macrolides, the bacterial number of P. aeruginosa PAO1 was found to be decreased in a dose-dependent manner by EM, CAM, or JM (Fig. 1B, D and F). The bacterial number of P. aeruginosa PAO1 treated with an NO donor was not decreased by GM (Fig. 1H). In treatment with nitrite, sensitivity of P. aeruginosa PAO1 was increased by the 14-membered macrolides EM and CAM but not by the 16-membered macrolide JM. This result was consistent with data on biofilm formation (18), modulation of c-di-GMP intracellular signaling (19), serum sensitivity (20), protein synthesis (21), and expression of flagellin (22). However, using an NO donor as a source of NO, the sensitivity also increased dose dependently by treatment with JM. Because the biofilm serves as a diffusion barrier to limit the access of many small molecules, including antibiotics, sub-MICs of the 14-membered macrolides may be clinically useful for the treatment of patients with P. aeruginosa infection through the repression of biofilm formation. We showed that 10 μg/ml of macrolides effectively increased the sensitivity of P. aeruginosa to NO. This concentration may be clinically achievable, because it has already been reported that the concentrations of azithromycin in alveolar macrophages and CAM in lung epithelial lining fluid reached 23 μg/ml (23) and 39.6 ± 41.1 μg/ml (24), respectively.

To investigate whether the increase in macrolide exposure-dependent sensitivity to exogenous nitrosative stress is a typical characteristic of P. aeruginosa, we also examined the bacterial numbers of other P. aeruginosa strains upon treatment with the NO donor in the absence and presence of EM (10 μg/ml). The MICs of EM for the 13 non-MDRP strains were all 256 or >256 μg/ml. In the absence and presence of EM, the average bacterial numbers of 13 non-MDRP strains, including PAO1, treated with an NO donor were 4.7 × 10⁵ ± 1.4 × 10⁵ and 5.1 × 10⁴ ± 2.3 × 10⁴ CFU/100 μl, respectively (Fig. 2A). The effect on bacterial numbers of non-MDRP strains treated with both NO and EM was strain dependent. Among the 13 non-MDRP strains analyzed, 9 became significantly more sensitive to nitrosative stress by exposure to EM (Fig. 2B). Furthermore, to investigate whether macrolide can increase the sensitivity of MDRP to exogenous nitrosative stress, we examined the bacterial numbers of MDRP strains treated with an NO donor in presence or absence of EM. Strains that were resistant to imipenem (IPM), fluoroquinolones (ciprofloxacin [CPFX] or levofloxacin [LVFX]), and amikacin (AMK) were defined as MDRP strains. MIC breakpoints for IPM, CPFX, LVFX, and AMK were ≥16, ≥4, ≥8, and ≥32 μg/ml, respectively. The MICs of EM for 30 of the 34 MDRP strains were 256 or >256 μg/ml. The MICs of EM for GUPR813, GUPR828, GUPR840, and GUPR848 were 64 μg/ml. In the absence and presence of EM, the average bacterial numbers of the 34 MDRP strains treated with an NO donor were 1.5 × 10⁵ ± 3.1 × 10⁴ and 1.1 × 10⁵ ± 3.9 × 10⁴ CFU/100 μl, respectively (Fig. 3A). Because the bacterial numbers of MDRP strains treated with an NO donor that were exposed to the sub-MIC of EM were strain dependent (Fig. 3C and D), we determined the sequence types (STs) of these 34 MDRP strains using whole-genome sequencing analysis (accession number DRA009821) and multilocus sequence typing (MLST) analysis (25, 26). Most of the MDRP strains (79.4% [27/34]) were high-risk ST235 clones (Fig. 3C and D). Therefore, to investigate whether macrolide can increase the sensitivity of MDRP ST235 to nitrosative stress, we analyzed the bacterial numbers of MDRP ST235 treated with NO in the absence or presence of EM. Under both conditions, average bacterial numbers of 27 MDRP ST235 strains were 1.4 × 10⁵ ± 3.5 × 10⁴ and 4.6 × 10⁴ ± 1.4 × 10⁴ CFU/100 μl, respectively, following treatment with an NO donor (Fig. 3B). Among the 27 MDRP ST235 strains analyzed, 11 became significantly more sensitive to the nitrosative stress via exposure to EM (Fig. 3C and D). However, the 7 MDRP non-ST235 strains did not become significantly more sensitive (Fig. 3C and D). Although 10 MDRP strains were highly sensitive to EDTA/NO alone, without EM (10₄ > CFU/100 μl), 7 of 10 MDRP strains were derived from urinary tract infection.

FIG 2.

Effect of EM on the sensitivity of the non-MDRP strains to exogenous nitrosative stress. The non-MDRP strains were grown in LB broth in the presence or absence of EM (10 μg/ml) at 37°C for 18 h (●). For the exogenous nitrosative stress assay, the 18-h grown non-MDRP strains were incubated in a saline solution containing 0.1% glucose, 10 mM MOPS (pH 7.0), and in the presence or absence of EM (10 μg/ml) with (■) or without (▲) 100 μM DETA/NO at 37°C for 22 h. The number of bacteria was determined using bacterial plate counts (CFU). All assays were repeated three times independently. The results are expressed as means ± standard errors. An unpaired t test was used to determine significant differences. *, P < 0.05. (A) Average numbers of bacteria in the non-MDRP group (n = 13). (B) Bacterial numbers of 13 non-MDRP strains.

FIG 3.

Effect of EM on the sensitivity of MDRP strains to exogenous nitrosative stress. MDRP strains were grown in LB broth in the presence or absence of EM (10 μg/ml) at 37°C for 18 h (●). For the exogenous nitrosative stress assay, the 18-h grown MDRP strains were incubated in a saline solution containing 0.1% glucose, 10 mM MOPS (pH 7.0), and in the presence or absence of EM (10 μg/ml) with (■) or without (▲) 100 μM DETA/NO at 37°C for 22 h. The number of bacteria was determined using bacterial plate counts (CFU). All assays were repeated more than three times independently. The results are expressed as means ± standard errors. An unpaired t test was used to determine significant differences. *, P < 0.05. (A) Average numbers of bacteria in the MDRP group (n = 34). (B) Average numbers of bacteria in the ST235 strains (n = 27). (C and D) Bacterial numbers of 34 MDRP strains.

In this study, we showed that one of the modes of action of low-dose macrolide treatment is increasing the sensitivity of P. aeruginosa to endogenous and exogenous nitrosative stresses. The survival of P. aeruginosa against the host immune system might be affected by sub-MIC macrolide treatment; hence, the effect of sub-MIC macrolide treatment on the sensitivity of P. aeruginosa to NO may partly explain the positive effects of long-term low-dose macrolide therapy. In fact, long-term, low-dose macrolide therapy has been reported to improve clinical outcomes in chronic airway diseases (2–4). However, treatment for 6 weeks may be too short to have a clinical effect (2). One of the reasons for the necessity of long-term treatment might be that there are some P. aeruginosa strains that were not significantly more susceptible to nitrosative stress upon treatment of airway conditions with macrolides.