Susceptibility of rapidly growing mycobacteria isolated from Australian cats to ivermectin,moxidectin,ceftiofur and florfenicol

Abstract

Objectives

Rapidly growing mycobacteria (RGM) infections in cats typically manifest as a panniculitis, requiring long-term antimicrobial therapy for resolution. The search for novel antimicrobial therapies to reduce treatment duration and improve the rate of clinical resolution is imperative. Accordingly, RGM isolates underwent susceptibility testing to some avermectins and other antibacterial drugs currently available.

Methods

Five Mycobacterium fortuitum and six Mycobacterium smegmatis isolates obtained from Australian cats underwent susceptibility testing by microbroth dilution to ivermectin, moxidectin, ceftiofur and florfenicol.

Results

All isolates were resistant to the highest concentrations of ivermectin, moxidectin and ceftiofur tested (1024 µg/ml, 256 μg/ml and 32 μg/ml, respectively). All isolates of M fortuitum were resistant to the highest concentration of florfenicol tested (128 µg/ml). The minimum inhibitory concentration range of florfenicol that inhibited growth of M smegmatis isolates was 32–64 µg/ml.

Conclusions and relevance

All drugs appear to have no efficacy in vitro for the treatment of RGM infections.

Short Communication

Rapidly growing mycobacteria (RGM) are capable of infecting a range of host species, including humans, domestic animals and wildlife.^1–3 They are acid-fast bacilli that produce visible colonies on routine media within 5–7 days when cultured at room temperature. ² They are divided into six phylogenic clusters, ⁴ with Mycobacterium fortuitum and Mycobacterium smegmatis the predominant species affecting cats in Australia. ² RGM have tropism for adipose tissue, ⁵ and in immunocompetent cats infections typically manifest as a panniculitis, introduced to the subcutis through catfight injuries or other breaches in the integument. ²

Treatment of RGM infections can be challenging, involving surgical debulking followed by prolonged administration of antimicrobial agents.^2,6,7 A combination of antibiotics is frequently administered, ⁸ with reports of antimicrobial therapy duration ranging from 3–60 months.^2,7,8 Doxycycline and fluoroquinolones are considered the initial agents of choice for treating this disease.^1,9 While isolates of both M fortuitum and M smegmatis have shown susceptibility to fluoroquinolones, some M smegmatis isolates have proven resistant to enrofloxacin and/or ciprofloxacin. ¹⁰ The presence of resistance coupled with daily dosing, prolonged administration and low clinical resolution drives the search for alternative antimicrobial therapies.

Recently, Lim et al ¹¹ tested the antimicrobial effectiveness of avermectins against different mycobacterial species. All avermectins tested showed promising bactericidal activity against multidrug resistant strains of Mycobacterium tuberculosis. Ivermectin and moxidectin inhibited the growth of a single M smegmatis isolate in vitro at 8 µg/ml and 4 μg/ml, respectively, suggesting that these agents may have a place in treating RGM infection in cats. Inhibitory effects of avermectins were not observed against Staphylococcus aureus, Escherichia coli or Pseudomonas aeruginosa, suggesting that mycobacteria may have a specific susceptibility to the avermectins.

The aim of this investigation was to determine the minimum inhibitory concentration (MIC) of ivermectin and moxidectin to inhibit growth of RGM obtained from Australian cats to ascertain the potential efficacy of these drugs as alternative therapies to treat clinical cases. As this study required reviving archived RGM isolates, it was considered opportunistic to also undertake susceptibility testing of isolates to other veterinary antimicrobials such as ceftiofur and florfenicol in order to explore alternative therapeutics for this chronic infection.

Materials and methods

Bacterial isolates

Feline RGM isolates from clinical cases encountered over the period 1983–2009 were available from the freeze-dried collection of the Veterinary Pathology Diagnostic Services, Faculty of Veterinary Science, The University of Sydney. All isolates had previously been identified to belong to a specific RGM species by a range of standard biochemical tests that included degree of acid-fastness, colony morphology, presence of nitrate reductase activity and presence of arylsulphatase activity, as well as via molecular techniques.^10,12 The biochemical identification of each isolate is described elsewhere, along with a description of preparation of the DNA template, amplification of 16S rRNA and heat shock protein loci, and DNA sequence. ¹⁰

MIC testing

MICs were determined using the microbroth dilution method in accordance with Clinical and Laboratory Standards Institute (CLSI) standard M24-A2. ¹³ Ivermectin and moxidectin were solubilised in 1 ml acetonitrile, water and hexane at a ratio of 4:1:5, respectively, ¹⁴ and florfenicol and ceftiofur were solubilised in 1 ml water prior to being diluted in Mueller Hinton broth.

Eleven two-fold serial dilutions of ivermectin, moxidectin, ceftiofur and florfenicol (all obtained from Sigma-Aldrich) were pipetted into microtitre wells (Nunc 163320). Dilutions of ivermectin and moxidectin were titrated between 1–1024 µg/ml and 0.25–256.00 µg/ml, respectively, to encompass a range either side of the MIC of 8 µg/ml and 4 μg/ml reported by Lim et al. ¹¹ Dilutions of ceftiofur ranged from 0.03–32.00 µg/ml and florfenicol ranged from 0.03–128.00 µg/ml.

Recommended quality control (QC) strains S aureus American Type Culture Collection (ATCC) 29213 and Mycobacterium peregrinum ATCC 700686 were tested concurrently to ciprofloxacin (dilutions between 0.03 and 32.00 µg/ml) in accordance with M24-A2. ¹³ Un-inoculated broth was used during each round of testing for detection of any contamination. All isolates including the QC strains also underwent susceptibility testing to the avermectin solvent by itself.

The MIC required to inhibit growth of S aureus ATCC 29213 was determined 24 h after incubation at 37^ºC, while MICs for RGM were determined after 72 h at 30^ºC. ¹³ For all drugs and isolates tested, the MIC was recorded as the lowest concentration of a drug that inhibited visible growth. If any inhibition of visible growth occurred, three further replications were performed at a higher concentration range for validation.

Results

Eleven isolates were available for this study. Five isolates were from the M fortuitum cluster and six isolates were from the M smegmatis cluster.

All isolates including the QC strains grew in the avermectin solvent. All M fortuitum and M smegmatis isolates grew in the highest concentrations of ivermectin, moxidectin and ceftiofur tested, being 1024 µg/ml, 256 μg/ml and 32 μg/ml, respectively.

Inhibition of visible growth occurred with initial florfenicol testing (dilution range 0.03–32.00 µg/ml) and, as a result, three subsequent assays were performed using a higher dilution range (0.125–128.000 µg/ml). All M fortuitum isolates grew in all dilutions of florfenicol on all three assays. The MIC₉₀ and MIC range of florfenicol to inhibit the growth of M smegmatis were 64 µg/ml and 32 to 64 µg/ml, respectively.

Ciprofloxacin inhibited the growth of the QC strains S aureus ATCC 29213 and M peregrinum ATCC 700686 within the expected ranges of 0.12–0.50 µg/ml and ⩽0.12–0.50 µg/ml, respectively, for all drug assays (seven assays were performed, one for each of the drugs tested and three subsequent florfenicol assays). Consistent endpoints were achieved with M peregrinum testing. Staphylococcus aureus varied by one two-fold dilution, which is within acceptable limits according to M24-A2. ¹³ No growth occurred in un-inoculated broth during testing.

Discussion

The exact mechanisms whereby RGM are resistant to avermectins, as observed here, are unknown. Natural drug resistance of Mycobacteria species is owing, in some part, to their lipid-rich cell wall, forming a barrier to the penetration of non-lipophilic antimicrobials. ¹⁵ However, both ivermectin and moxidectin are highly lipophilic,^16,17 and theoretically capable of penetrating the mycobacterial cell wall. Thus, the cell wall barrier alone does not explain the high degree of drug resistance seen in some Mycobacteria species, ¹⁸ suggesting contributions by one or more other mechanisms.

The differing MIC result between our study and that of Lim et al ¹¹ may be owing to the strain of M smegmatis tested, as only one isolate of M smegmatis was tested in their study. Additionally, the differing MICs could be due to the testing procedure itself. Lim et al ¹¹ used the bacterial growth indicator 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide, whereas the microbroth dilution method was used in our study. The microbroth dilution method is the gold standard for susceptibility testing of RGM, ¹⁹ and other susceptibility tests are considered to provide more variable results.^19–22 Microbroth dilution is not without limitations, however, with the precise endpoint often difficult to read. ²³ Further attempts to replicate the data of Lim et al ¹¹ with respect to ivermectin, and M tuberculosis by Muhammed Ameen and Drancourt ²⁴ using the agar dilution method, generated different MICs, suggesting that determination of the antimycobacterial activity of avermectins may be affected by the assay method.

Plasma drug concentrations achievable in patients must exceed the MIC. In pharmacokinetic studies of ivermectin in cats, peak serum concentrations (C_max) after receiving a single subcutaneous injection of 0.2 mg/kg ivermectin averaged 16.75 ng/ml, ¹⁷ which is several fold lower than the MIC of 8 µg/ml reported by Lim et al. ¹¹ Similarly, while there are no pharmacokinetic studies of moxidectin in cats, the C_max of moxidectin in dogs and horses receiving standard doses have been obtained at 234 ng/ml ¹⁶ and 70.3 ng/ml, ²⁵ respectively, and are several fold lower than the MIC of 4 µg/ml obtained by Lim et al, ¹¹ providing further evidence that ivermectin and moxidectin are unlikely efficacious against in vivo M fortuitum or M smegmatis infections.

Not surprisingly, RGM were resistant to ceftiofur, which is comparable with other studies recognising RGM resistance to β-lactams. Other ‘third-generation’ cephalosporins such as cefoperazone, ceftazidime and cefovecin have proven ineffective in RGM susceptibility studies.^26–28 Penicillins have similarly been demonstrated to be ineffective in the treatment of mycobacterial infections, likely owing to β-lactamase activity present in all mycobacteria. ²⁷ Several studies of β-lactamases produced by M fortuitum found that the level of these enzymes correlated inversely with their susceptibility to β-lactams.^26,29,30 Similar results have also been found with the M smegmatis cluster. ²⁷ A gene encoding a cephalosporin-specific β-lactamase (cephalosporinase) has been identified in the M smegmatis genome, ²⁷ and may account for the resistance seen against ceftiofur in this study. The aforementioned lipid-rich cell wall of mycobacteria is another plausible explanation for resistance given the low lipophilicity of ceftiofur. ³¹

While florfenicol is not currently approved for use in companion animals, ³¹ its broad spectrum and greater in vitro activity against pathogenic bacteria compared with chloramphenicol and thiamphenicol may have future applications as a treatment option for cats. ³² However, M fortuitum isolates showed no evidence of susceptibility to florfenicol and the MIC range of 32–64 µg/ml to inhibit the M smegmatis cluster is likely too high to be of any therapeutic benefit. While no pharmacokinetic studies of florfenicol usage in cats are available, canine studies demonstrated a plasma C_max of 6.18 µg/ml after a dose of 20 mg/kg, ³³ which is several fold lower than the MIC range obtained here. Similarly, given the tropism of RGM for adipose tissue, sustainable concentrations of florfenicol in fat would need to be attained for suitable therapy. However, florfenicol does not penetrate adipose tissue easily.^34,35

The difference in susceptibility to florfenicol between M smegmatis and M fortuitum in this study is likely owing to a structural difference in the two bacteria or through the dissemination of resistance genes in the M fortuitum cluster. ³⁶ This difference parallels the Australian study by Malik et al, ² in which M fortuitum strains were more commonly resistant to multiple antibiotics than other clusters. However, other studies have found M fortuitum to be more susceptible in vitro to antimicrobials than other RGM species, ⁷ suggesting possible differing geographical susceptibility of the M fortuitum cluster. The results also suggest that exposure to high concentrations of florfenicol could be a useful feature with which to distinguish between M fortuitum and M smegmatis clusters in vitro.

CLSI standard M24-A2 provides no information on susceptibility testing of RGM to any of the antimicrobial drugs used in this study, ¹³ and therefore, as the susceptibility breakpoints of the two QC strains to this drug are available, ciprofloxacin assays were run concurrently. The fact that ciprofloxacin inhibited the growth of the QC strains within their expected MIC ranges and that no growth occurred in the un-inoculated broth during each round of testing highlights that the testing method was appropriate and there was no evidence of contamination during susceptibility testing.

Conclusions

While there are discrepancies between mycobacterial drug susceptibility determined in vitro with the activity of the drug observed in vivo, ³⁷ in vitro susceptibility testing indicates that ivermectin, moxidectin, ceftiofur and florfenicol are unlikely to have any efficacy in treating RGM infections in Australian cats.