Biofilms of Mycobacterium abscessus Complex Can Be Sensitized to Antibiotics by Disaggregation and Oxygenation

Abstract

Pulmonary infection with the multidrug-resistant Mycobacterium abscessus complex (MABSC) is difficult to treat in individuals with cystic fibrosis (CF). MABSC grows as biofilm aggregates in CF patient lungs, which are known to have anaerobic niches. How aggregation and anoxic conditions affect antibiotic tolerance is not well understood. We sought to determine whether disaggregation and oxygen availability sensitize MABSC isolates to recommended antibiotics. We tested the susceptibilities of 33 isolates from 22 CF patients with MABSC infection and a reference strain to the following antibiotics: amikacin, azithromycin, cefoxitin, ciprofloxacin, clarithromycin, imipenem, kanamycin, linezolid, moxifloxacin, rifampin, tigecycline, and sulfamethoxazole-trimethoprim. Isolates were grown in Mueller-Hinton broth with and without the disaggregating detergent Tween 80 (5%). Time-kill curves at days 1 and 3 were generated for oxic and anoxic amikacin treatment in 4-fold dilutions ranging from 2 to 512 mg liter^-1 Scanning electron microscopy was used to visualize the aggregation patterns, while confocal laser scanning microscopy and microrespirometry were used to visualize biofilm growth patterns. Disruption of MABSC aggregates increased susceptibility to amikacin, tigecycline, kanamycin, azithromycin, imipenem, cefoxitin, and clarithromycin (P < 0.05, n = 29 to 31). Oxygenation enhanced the killing of disaggregated MABSC isolates by amikacin (P < 0.05) by 1 to 6 log units when 2 to 512 mg liter^-1 of amikacin was used. This study explains why current drug susceptibility testing results correlate poorly with treatment outcomes. The conditions achieved by oxic culturing of planktonic isolates in vitro do not resemble the hypoxic conditions in CF patient lungs. Biofilm disruption and increased O₂ availability during antibiotic therapy may be new therapeutic strategies for chronic MABSC infection.

Keywords: Mycobacterium abscessus complex; antimicrobial resistance; biofilm; cystic fibrosis; oxygenation.

FIG 1

Aggregation and cording formation by MABCS with rough morphotypes. (A) Cord formation of MABSC in Ziehl-Neelsen-stained CF sputum sample. (B) MABSC aggregate with cells of the rough morphotype isolated from a CF patient sputum sample visualized by scanning electron microscopy (SEM). (C and D) Enlarged images of organized cord formation in the MABSC aggregate with cells of the rough morphotype (C) and smooth morphotype (D) isolated from a CF patient sputum sample visualized by SEM.

FIG 2

The growth of MABSC is influenced by disaggregation. The images show the effect of Tween 80, which breaks up MABSC isolates from liquid cultures. Bacterial aggregation at four different Tween 80 concentrations is shown. Samples were stained with Syto 9, and the images were obtained using a 63× (numerical aperture [NA], 1.4) Zeiss objective on a Zeiss 880 CLSM. Bars, 20 μm at a magnification of ×630. Isolates I to III were isolated from three CF patients +10, +3, and +5 years after collection of the first sample positive for M. abscessus.

FIG 3

The size of MABSC aggregates from liquid cultures decreases as the concentration of Tween 80 increases. The fractions of large (>10 μm³; white) versus small (1 to 10 μm³; black) aggregates of the MABSC ATCC 19977 smooth reference strain are shown.

FIG 4

Disaggregation with Tween 80 sensitizes MABSC to antimycobacterial drugs. The effect of 5% Tween 80 on MABSC isolates (n = 31) was tested by disk diffusion susceptibility testing using amikacin, azithromycin, cefoxitin, ciprofloxacin, clarithromycin, imipenem, kanamycin, linezolid, moxifloxacin, rifampin, tigecycline, and sulfamethoxazole-trimethoprim. Aggregated MABSC isolates were grown and incubated in MH and plated by normal plate spreading on blood agar plates. Disaggregated MABSC isolates were suspended in MH plus 5% Tween 80, incubated for 2 to 5 days, and subsequently added onto blood agar plates and uniformly distributed by gently tipping the plates. Excess liquid was discharged. All isolates that were completely resistant regardless of Tween 80 are represented by the flat red line on the x axis. Statistical significance was determined using a paired t test (P ≤ 0.05). NS, not significant.

FIG 5

Multidrug resistance to antimycobacterial drugs in MABSC is lowered by disaggregation of the isolates with Tween 80. MABSC isolates (n = 31) were tested by disk diffusion susceptibility testing, using (per disk) 30 μg amikacin, 15 μg azithromycin, 30 μg cefoxitin, 5 μg ciprofloxacin, 15 μg clarithromycin, 10 μg imipenem, 30 μg kanamycin, 10/30 μg linezolid, 5 μg moxifloxacin, 5 μg rifampin, 15 μg tigecycline, and 1.25 μg/23.75 μg sulfamethoxazole-trimethoprim. Isolates with zone diameters of 0 mm were considered resistant. Aggregated MABSC isolates were grown and incubated in MH and plated by normal plate spreading on blood agar plates. Disaggregated MABSC isolates were suspended in MH plus 5% Tween 80, incubated for 2 to 5 days, and subsequently added onto blood agar plates and uniformly distributed by gently tipping the plates. Excess liquid was discharged. Statistical significance was determined using a paired t test (P ≤ 0.05).

FIG 6

The oxygen consumption of MABSC is influenced by the degree of bacterial aggregation. The graphs show the microrespirometry of O₂ in aerobic cultures with smooth and rough CF MABSC isolates and the ATCC 19977 strain during different stages of bacterial disaggregation, as determined by the Tween 80 concentrations: 0% (black), 1.25% (red), 2.5% (green), 5% (blue). Isolates I to III were isolated from three CF patients +10, +3, and +5 years after the first sample was positive for MABSC. (A to G) The oxygen consumption with 5% Tween 80 was significantly higher than the O₂ consumption with 0% Tween 80 for all isolates (P < 0.0001). Error bars indicate the mean ± standard error of the mean (n = 3). Statistical significance was assessed using two-way ANOVA, followed by Bonferroni multiple-comparison tests. (H to N) Effect of the Tween 80 dilution on the numbers of CFU per milliliter. Error bars indicate the mean ± standard error of the mean (n = 3). Statistical significance was determined using one-way ANOVA, followed by Dunnett's multiple-comparison test (*, P ≤ 0.05).

FIG 7

Aerobic conditions increase the bacterial killing of M. abscessus complex isolates during amikacin treatment. Isolates from three CF patients were cultured for 0, 1, or 3 days with amikacin, as indicated. On each day, the number of CFU per milliliter was recorded for each amikacin treatment for triplicate individual experiments. Killing was calculated as the log reduction relative to the number of CFU per milliliter of untreated samples on each day. Isolates I to III were from three CF patients +10, +3, and +5 years after first sample was positive for M. abscessus. (A) Effect of increasing amikacin concentrations ranging from 0 to 512 μg ml⁻¹ on smooth CF isolates I^SM+10, II^SM+3, and III^SM+5 incubated for 0, 1, and 3 days under oxic (red) and anoxic (black) conditions. (B) Effect of increasing amikacin concentrations ranging from 0 to 512 μg ml⁻¹ on rough CF isolates I^RG+10, II^RG+3, and III^RG+5 incubated for 0, 1, and 3 days under oxic (red) and anoxic (black) conditions. Error bars indicate the mean ± standard error of the mean (n = 3). The statistical significance of the difference between oxic and anoxic treatments was assessed using a two-way ANOVA, followed by Bonferroni multiple-comparison tests (P ≤ 0.05).