Introduction

Buruli ulcer (BU) is a neglected tropical disease that presents as skin nodules, plaques or oedematous lesions that can progress to open ulcers [1]. BU is caused by infection with Mycobacterium ulcerans, a mycobacterium that is related to the causative agents of tuberculosis and leprosy [2]. Most BU patients are children under the age of 15 [3]. The mode of transmission of the disease is not well understood. Superstitious beliefs predominate in rural West Africa, resulting in delayed treatment seeking and stigmatization of patients [4,5]. Mortality associated with BU is low nevertheless morbidity is high. Extensive ulcers frequently lead to lifelong physical disability [6,7].

No vaccine is available against Buruli ulcer and management focuses on early case detection and treatment with surgery and antibiotics [8]. Previously, Buruli ulcer was treated with surgical excision only, but since 2004, an 8-week course of rifampicin and streptomycin is the standard treatment [9–11].

In case patients report early with limited lesions, the 8-week course of rifampicin and streptomycin delivers a good quality of life at long-term follow up [12]. However, the median time to heal is still 18 to 30 weeks, depending on the size of the lesion [10].

Patients presenting late to health care facilities have a much poorer prognosis and many suffer from functional limitations due to the disease [6,7].

There is also the issue that daily injections with streptomycin are impractical and can have serious side effects, such as ototoxicity [13]. It has been shown that all-oral treatment with rifampicin and a macrolide or quinolone results in high cure rates [14]. Shorter duration of antibiotic courses and more safe treatment regimes that reduce the time to healing are desirable for Buruli ulcer.

Avermectins are macrolides that are used to treat helminth-infection (such as strongyloidiasis or onchocerciasis) and parasitic infection (scabies) in humans and in animals. These orally administered drugs are well tolerated and available worldwide. Ivermectin is on the essential drugs list of the World Health Organization. A recent report showed that avermectins, including ivermectin, moxidectin and selamectin, inhibit the growth of different M. tuberculosis strains in-vitro at concentrations of 2 to 8 μg/ml [15]. Motivated by these findings, we tested if avermectins also inhibit and kill M. ulcerans.

Materials and Methods

Results

M. ulcerans growth is inhibited by avermectins

M. ulcerans JKD8049, M. ulcerans 1117–13 and M. marinum were grown in MGIT tubes in the presence of increasing concentrations of the ivermectin and moxidectin. At three weeks, no fluorescence was observed at 8 μg/ml of ivermectin for M. ulcerans JKD8049 and at 4 μg/ml for M. ulcerans 1117–13 (Table 1). Moxidectin inhibited the growth of M. ulcerans JKD8049 at 4 μg/ml. The MIC for M. marinum was 32 μg/ml for ivermectin and above 64 μg/ml for moxidectin (Table 1). These data show that M. ulcerans but not M. marinum is susceptible to avermectins. Growth of all bacteria was observed in the control tubes containing only the solvent and no growth was observed in the tubes containing 0.1 μl/ml rifampicin.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Minimal inhibitory concentration (MIC) of avermectins compared with rifampicin.

https://doi.org/10.1371/journal.pntd.0003549.t001

M. ulcerans killing by avermectins

Time-kill assays were then performed to assess if avermectins not only inhibit M. ulcerans growth but also kill the bacterium. Based on the MIC results above, a low and high dose of ivermectin was tested and M. ulcerans JKD8049 was exposed to either 8 μg/ml or 20 μg/ml ivermectin for eight weeks. A dose-dependent killing effect was observed with no CFU detected from week-4 onwards at 20 μg/ml and week-5 onwards for 8 μg/ml (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Time-kill assay showing the effect of ivermectin against M. ulcerans over 8-weeks.

M. ulcerans was grown in 25cm² culture flasks in the presence of 0, 8 and 20μg/ml ivermectin. Ethanol (solvent for ivermectin) at the concentration in the IVM-20 dose was used in the no-drug control. Weekly, ten-fold dilutions of each culture were plated onto 7H10 agar, using a 3-μl spot-dilution method, with five replicates per dilution. The plates were examined for growth after incubation for 8-weeks at 30°C, and colony forming units per ml calculated. The mean and range for duplicate biological experiments are shown.

https://doi.org/10.1371/journal.pntd.0003549.g001

Bioluminescence is an ATP-dependent process and it is therefore an excellent dynamic reporter of cellular metabolic activity [17]. A time-kill experiment was thus performed using bioluminescent M. ulcerans. Although the time frame of this experiment was quite short (21 hours), continuous monitoring over that period again showed a dose-dependent impact of ivermectin on bacterial viability (Fig 2). The reduction in bioluminescence at 21 hours was higher in ivermectin at all concentrations tested (8, 16 and 32 μg/ml) compared to the rifampicin positive control (Fig 2). Interestingly, the greatest impact on bacterial viability was observed in the presence of a combined dose of 8 μg/ml ivermectin and 0.1 μg/ml rifampicin (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Dose-dependent change in relative light units (RLU) over 21 hours for bioluminescent M. ulcerans exposed to different concentrations and combinations of ivermectin and rifampicin.

Ethanol (solvent for ivermectin) at the concentration in the IVM-32 dose was used in the no-drug control (EtOH). Depicted is the mean of at least technical triplicates for measurements made every 300 seconds over 21 hours.

https://doi.org/10.1371/journal.pntd.0003549.g002

Discussion

The current treatment regimen of 8-weeks streptomycin and rifampicin for Buruli ulcer is highly effective but also problematic and promising alternative all-oral treatments are sought [13,14]. Here we have shown that ivermectin and moxidectin were able to inhibit the growth of different M. ulcerans strains at 4–8 μg/ml. These findings are comparable to previous research where MICs of 1–8 μg/ml against M. tuberculosis with ivermectin, selamectin, moxidectin or doramectin were reported [15].

M. marinum, despite its close relationship to M. ulcerans, showed low susceptibility to both ivermectin and moxidectin. We speculate that this difference may be explained at least in part by the abundance of intact transporters and efflux systems in M. marinum and a corresponding scarcity of the equivalent systems in M. ulcerans [18,19].

The 8-week time-kill experiment (Fig 1) showed that ivermectin at concentrations of 8 μg/ml and above has a likely bactericidal effect on M. ulcerans. Future experiments will test lower drug concentrations and use higher starting doses of bacteria. The long doubling time of M. ulcerans (>48h) complicates laboratory-testing methods using traditional culture and colony counting approaches. Here we used bioluminescence as a rapid read-out of cell viability and found it a useful technique for examining these slow-growing mycobacteria (Fig 2). Given the MIC findings (Table 1), we were surprised to observe in the bioluminescent time-kill experiments that ivermectin performed better than rifampicin at their given MIC concentrations. Further investigation of this difference is warranted, probably by conducting the experiments over an extended time period and beyond the 21 hours we were able to achieve here. A combination of 0.1 μg/ml rifampicin and 8 μg/ml ivermectin showed a synergistic killing effect. Clinically, rifampicin may however decrease the serum concentration of ivermectin by the induction of P-glycoprotein/ABCB1 in humans [20]. The extent to which this might be the case is not known and requires further testing.

In the context of avermectins to treat M. tuberculosis infections, it has been argued that MICs of 1–8 μg/ml are not achievable in humans [21]. Avermectins are mostly administered in much lower doses and in single-doses for the treatment of helminth infection. A maximum plasma concentration of 54.4 ng/ml was observed in healthy volunteers receiving 150 μg/kg ivermectin [22].

There are few public data on the safety of higher doses or prolonged courses of avermectins in humans and it unclear at what doses and over what time avermectins would have to be given to Buruli ulcer patients. We would argue that obtaining avermectin plasma concentrations at concentrations that approach MICs derived from in vitro experiments might not be needed for clinical efficacy. In our bioluminescent time-kill assay, 8 μg/ml ivermectin was superior to 0.1 μg/ml rifampicin. It may be that the treatment duration for BU can be reduced and that complete eradication of the microbe is not needed. Interference with the M. ulcerans mycolactone toxin synthesis machinery at concentrations that are sub-inhibitory for growth might be sufficient to permit host immunity to then clear the infection.

Acknowledgments

We thank Estelle Marion for provision of M. ulcerans 1117–13, Siouxsie Wiles for the provision of the bioluminescent reporter plasmid and Renee Marcisin for preparation of the bioluminescent version of M. ulcerans JKD8049.