Antimicrobial peptides against drug resistant Mycobacterium abscessus

Abstract

Mycobacterium abscessus (MAB) comprise rapidly growing, often multidrug-resistant (MDR), nontuberculous mycobacteria responsible for pulmonary and other infections in susceptible hosts. Antimicrobial peptides (APs) are natural and synthetic antimicrobials active against a range of microorganisms including mycobacteria. We evaluated APs activity against MAB reference and clinical strains. We observed minimal inhibitory concentrations of 1.6 to > 50 μg/ml. Further work with the most active AP demonstrated protection of Acanthamoeba castellanii (AC) from killing by ingested MAB including MDR MAB strains. Antimicrobial peptides offer an attractive potential option for treatment of drug resistant treatment-refractory MAB.

1. Introduction

Mycobacterium abscessus (MAB) encompasses a group of rapidly growing nontuberculous mycobacteria (NTM) currently divided into 3 highly related taxa: M. abscessus subsp. abscessus, M. abscessus subsp. massiliense, and M. abscessus subsp. bolletii [1]. MAB is most commonly implicated in hard-to-treat chronic pulmonary infections in susceptible hosts, especially those with chronic lung disease such as cystic fibrosis (CF), leading to high morbidity and mortality [2]. MAB pulmonary disease requires long treatment regimens with limited drug combination options and can lead to the development of drug resistance [3]. Despite the clear need for new therapeutic strategies, there are currently only few compounds focused on MAB in clinical trials [4]. Synthetic short cationic antimicrobial peptides (APs) have been shown to have good activity against various bacteria and fungi. These agents have been assayed against Mycobacterium tuberculosis with promising killing activity [5] and low emergence of drug resistance. Moreover, they show synergistic activity when combined with drugs commonly used against mycobacteria [6].

Here, we evaluated if antimicrobial peptides (AP) show activity against MAB clinical isolates by measuring Minimal Inhibitory Concentration (MIC) and using Acanthamoeba castellani (A. castellani) - MAB infection assays. We show that antimicrobial peptides were active against a battery of reference and clinical MAB including a MDR MAB outbreak strain [7]. Moreover, AP activity protected A. castellani from killing by intracellular MAB virulent clinical strains.

2. Material and Methods

2.1. Mycobacteria

Mycobacteria were grown in Middlebrook 7H9 broth supplemented with 10 % ADC and 0.1% tyloxapol (Sigma Aldrich, St. Louis, MO, USA) to Optical density (OD) of 0.6. Mycobacteria were centrifuged, supernatant discarded, and the pellet suspended in 20 mL 7H9 broth. Solid glass beads (3mm diameter Sigma Aldrich, St. Louis, MO, USA) were added and samples homogenized by vortex for 2 min. Aliquots of 1-mL were transferred to 1.5-mL tubes and passed through a 26G needle for 10 times. Samples were then sonicated three times for 15 sec with 20 sec intervals in a bath sonicator and let stand at room temperature for 15 min. After adjustment to an OD₆₀₀ of 0.2, 500 μL aliquots were stored at −80 °C. Frozen stocks of mycobacterial cells were enumerated by plating 10-fold serial dilutions onto Middlebrook 7H10 Agar (Becton Dickinson, Franklin Lakes, NJ, USA).

We used three well-characterized M. abscessus subsp. massiliense strains serially collected from a CF patient: MAB_062600_1635 (= strain B1, early infection [10]), MAB_030804_1651 (= strain B5, mid-course infection [10]), and MAB_010708_1655 (= strain B8, late stage infection [10], = strain 2B-0107 [11]). These three clinical strains were transformed with pCherry-Hygromycin plasmid (Addgene) for mcherry red fluorescence visualization of M. abscessus infection in A. castellanii. In addition, 25 clinical M. abscessus strains including both rough and smooth phenotypes of M. abscessus subsp. abscessus and M. abscessus subsp. massiliense isolated from different CF and non-CF bronchiectasis patients were used in the study.

2.2. Antimicrobial Peptides:

Antimicrobial peptides (WatsonBio, Houston, TX, USA) were diluted into ultra pure water (KD Medical, Columbia, MD, USA) to a concentration of 4 mg/mL and 50- μL aliquots transferred to 1.5-mL polypropylene O-ring tubes (Sarstedt, Numbrecht, Germany). Antimicrobial peptides (APs) used in this work include: AP1 – ILPWKWRWWKWRR [5], AP2 - ILPWKWRWWKWWR (R12W substitution of AP1, this work), AP3 – D enantiomeric AP1, AP4 – ILRWKWRWWRWRR [5], AP5 – GLFDVIKKVASVIGGL [8] and AP6 - KRAKKFFKKLK (ATRA-1A) [9]

Clofazimine (Sigma Aldrich, St. Louis, MO, USA), used as a positive antimicrobial control for the serial M. abscessus subsp. massiliense strains, was diluted to a concentration of 2 mg/mL in DMSO. All stock solutions (APs and clofazimine) were kept at −80 °C until use.

2.3. Minimal Inhibitory Concentration (MIC) assays

MIC assays were performed in triplicate in polypropylene non-treated 96-well microplates (Corning, Corning, USA) with a modified (excluding magnesium sulfate) 7H9 broth as described previously [5]. Frozen stocks of mycobacteria were thawed and diluted to approximately 5 × 10⁵ CFU/mL in modified 7H9. APs and clofazimine concentration range tested was 50 to 0.2 μg/mL. Control wells included uninoculated broth, peptides or mycobacteria in broth. Plates were incubated at 30°C for 3 days. Minimal inhibitory concentration was determined as the lowest concentration of drug that prevented visible growth.

2.3. Acanthamoeba castellanii

A. castellanii ATCC 30010 trophozoites were grown in 15 mL Peptone Yeast Extract Glucose broth (PYG) in 75 cm² cell culture flasks at 28.5 °C until confluency was reached. Cells were detached by gently tapping the flasks followed by centrifugation at 500 g for 10 min. Cells were diluted in PYG with 10% DMSO (v/v) to a final concentration of 10⁷ cells/mL and stored in liquid nitrogen. For each experiment, aliquots were thawed at 37 °C and used for up to 3 weeks or 4 passages.

2.4. Acanthamoeba castellanii – MAB Infection

A. castelanii ATCC 30010 trophozoites were transferred from cell culture flasks with PYG, to 24-well plates at ~10⁶ cells/mL and incubated at 28.5 °C overnight. Adherent amoebae were washed three times with PYG and 100 μL of mycobacteria at 10 ⁸ CFU/mL was added. Following infection for 3 h at 28.5 °C, amoebae were washed three times with PYG and 1 mL of fresh PYG was added (Fig 1A). Plates were used in AP treatment experiments and clofazimine controls.

Figure 1.

A: Protocol describing infection of A. castellanii ATCC 50370 by Mycobacterium abscessus. B. Fluorescent microscopic images demonstrating A. castellanii ATCC 30010 infected with MAB_062600_1635 (1), MAB_030804_1651 (2) and MAB_010708_1655 (3) expressing m-Cherry. The scale bars on lower left corner of the images panels correspond to 5 micrometer magnification.

2.5. Treatment of infected Acanthamoeba castellanii by antimicrobial peptide and clofazimine

Antimicrobial peptide at 3X MIC were added to infected amoebae (described in 2.4) and incubated for 48 h at 28.5 °C. After 48 h incubation 100 μl of the culture were removed and centrifuged at 500 g for 10 min and supernatant used for LDH measurements. LDH was measured using CytoTox 96® NonRadioactive Cytotoxicity Assay following manufacturer`s instructions. Protective effect by AP or clofazimine against M. abscessus-induced killing of A. castellanii (AC) was assessed using the following LDH release ratios: [LDH released from infected AC (with no AP or clofazimine) / LDH released from infected AC in the presence of APs or clofazimine]. A higher LDH ratio in the presence of APs or clofazimine is an indicator of greater amoeba survival and less toxicity to the amoeba from intracellular infection by MAB.

2.6. Fluorescence Imaging

Structured Illumination Microscope (SIM) images were acquired with a VT-iSIM Imaging System scanner (VisiTech International Sunderland, UK) on an Olympus IX 81 microscope using an Olympus UPLSAPO 100 × 1.49 NA Oil objective (Olympus America, Inc. Melville, NY) and dual Hamamatsu CMOS Orca-Flash 4 (Hamatsu USA Bridgewater, NJ) cameras. The total acquisition system was controlled using Metamorph software (Molecular Devices, LLC, San Jose, CA). The fluorescence from the fluorescent protein mCherry was excited with a 561nm laser and the emission light was filtered using a long pass 590nm emission filter before the camera. 3D volumes were taken at an interslice distance of 100nm for a total of 30–36 individual planes. Exposure time for each plane was 250ms. Images were deconvolved and computationally de-striped using an iSIM specific ImageJ plugin from the company (Microvolution LLC Cupertino, CA).

2.6. Statistical analysis

Results are shown as the mean ± S.D. (standard deviation) of three independent experiments. Significance (p-value < 0.05) was determined by paired Student’s t-test.

3. Results and discussion

In this work, we tested six antimicrobial peptides (APs) against drug susceptible and drug resistant reference and clinical strains of MAB. APs were chosen in part based on previously reported MIC against M. tuberculosis and Mycobacterium smegmatis [5, 9]. We first screened all 6 APs using three well-characterized M. abscessus subsp. massiliense strains serially collected from a CF patient. We observed relatively low MICs of AP1 against all three strains (MIC range 1.5–3.1 μg/mL, Table 1) including the MDR strain MAB_010708_1655 previously implicated in a seminal outbreak in a CF center in Seattle, Washington, USA [7, 11]. Similar MIC results were obtained with AP3 (the D enantiomeric AP1, data not shown). In contrast, AP2 which differs from AP1 by a single amino acid (R12W) showed MICs higher than 50 μg/mL (Table 1). We also observed similar MIC (3.1 μg/mL) of AP1 against laboratory reference strain M. abscessus ATCC 19977 (Table 1). The remaining APs 4, 5, and 6 also showed elevated MICs and were not further characterized.

Table 1.

Minimal Inhibitory Concentration (μg/mL) of Antimicrobial Peptides against M. abscessus strains

Strain	AP1	AP2	AP4	AP5	AP6
MAB ATCC 19977	3.1	>50	>50	>50	>50
MAB_062600_1635	1.6	>50	>50	>50	>50
MAB_030804_1651	1.6	>50	>50	>50	>50
MAB_010708_1655	1.6	>50	>50	>50	>50

Having shown promising activity of AP1 against well characterized MAB strains, we sought to test additional strains. AP1 showed low MICs against additional 25 clinical MAB with values ranging 1.5 to 6.2 μg/mL. As expected, AP2 showed MICs higher than 50 μg/mL for these strains (Data Not Shown).

In a previous study we showed strains MAB_062600_1635 and MAB_010708_1655 could infect A. castellanii with the latter strain being more cytotoxic [10]. We, therefore, decided to use this same host-pathogen model to confirm bacterial internalization via 3D structured illumination microscopy and assess AP1 activity using LDH release assays. All three strains were internalized by amoeba, as shown in Fig 1. (Panel B).

To study AP activity in the amoeba-mycobacteria model, we calculated the ratio of LDH released from MAB infected amoeba in the absence of AP (untreated) divided by LDH released from infected amoeba in the presence of AP (treated). Therefore, higher LDH ratios (greater than 1) indicate protection of amoeba by the drug against MAB-induced killing. Different concentrations of AP1 (1–3 × MIC) were tested in amoeba infected with MAB_062600_1635, MAB_030804_1651 or MAB_010708_1655 (Fig 2A, B and C, respectively). The antimycobacterial drug clofazimine was used as positive controls. AP1 at a concentration of 3 × MIC was effective at inhibiting MAB killing of amoeba for all 3 strains as evident by the high LDH release ratio (p <0.05) (Fig 2)). Higher concentrations were not tested due to peptide precipitation. Effects of AP1 at lower concentrations (1X, 2X MIC) were not statistically significant. Clofazimine (positive control) at a concentration of 3X MIC was effective in protecting against MAB_062600_1635, MAB_030804_1651, and MAB_010708_1655 killing of amoeba. Overall, our results show a significant antimicrobial activity of AP1 against MAB in the amoeba-mycobacteria infection model.

Figure 2.

LDH release ratio of LDH from Acanthamoeba castellanii infected with M. abscessus (untreated) to the LDH released from M. abscessus infected Acanthamoeba castellanii treated with antimicrobial peptide AP1 or Clofazimine CLO. Different M. abscessus strains used in the study are MAB_062600_1635 (A), MAB_030804_1651 (B) and MAB_010708_1655 (C).

Current treatments of MAB lung infection involve several drug combinations for a lengthy period of time [2] often resulting in unsuccessful eradication [12]. MAB intrinsic drug resistance, due to its low cell wall permeability combined with inducible and acquired mutational resistance to macrolides and aminoglycosides [4], make it a treatment challenge. With current suboptimal drug regimens for treatment of MAB disease, there is an urgent need for novel therapeutic approaches.

In the search for novel MAB treatment candidates, we hypothesized that antimicrobial peptides previously shown to be active against other mycobacterial species might be effective against MAB and could serve as a backbone for targeted development of new compounds. As a proof of principle, here we characterized a candidate (AP1) showing low MIC to multiple reference and clinical strains of MAB including MDR and outbreak strains. We also highlight that a single amino acid change with loss of one positively charged arginine was sufficient to abolish the antimicrobial activity of AP1.

To further characterize AP1 activity, we sought a suitable host-pathogen model. Free-living amoebae are natural hosts for mycobacteria and have been shown to be a good model for studying host-pathogen interactions as they display similarities to human macrophages [13, 14, 16]. We previously confirmed the suitability of the amoeba as a host model for MAB virulence using 2 MAB strains included in this study [10]. We then sought to expand the use of this amoeba model for assessing antimicrobial activity against MAB by measuring the ability of AP to prevent MAB-induced amoeba killing. Our data show that AP1 at a concentration of 3X MIC is active against MAB in the amoeba model. AP1 at these concentrations had no cytotoxic effect on A. castellani alone (data not shown). Notably, AP1 protected amoeba against killing by different strains including a highly virulent, multidrug resistant, outbreak-related strain. Designed antimicrobial peptides have gained attention to treat M. abscessus, especially multidrug resistant M. abscessus. Dermody et al. has recently described two synthetic antimicrobial peptides (RP557 and JC41–4 dAMPs) with promising killing activity against rough and smooth M. abscessus [15]

Having shown AP1 to be a promising candidate against MAB, future work will focus on 3 lines of research: 1) structure–activity relationship (SAR) studies to maximize AP1 antimicrobial activity with minimal host cytotoxicity, 2) assessment of antimicrobial activity using zebra fish and other animal models, 3) evaluation of peptide delivery options for optimal activity in the lung.