Cyclohexyl-griselimycin Is Active against Mycobacterium abscessus in Mice

ABSTRACT Cyclohexyl-griselimycin is a preclinical candidate for use against tuberculosis (TB). Here, we show that this oral cyclodepsipeptide is also active against the intrinsically drug-resistant nontuberculous mycobacterium Mycobacterium abscessus in vitro and in a mouse model of infection. This adds a novel advanced lead compound to the M. abscessus drug pipeline and supports a strategy of screening chemical matter generated in TB drug discovery efforts to fast-track the discovery of novel antibiotics against M. abscessus.

discovery in the 1960s. However, the first human studies were halted due to poor oral bioavailability (12,14,15). These forgotten natural products were recently revisited by investigators from Sanofi in association with TB Alliance to identify analogs with improved PK properties. The cyclohexyl analog CGM (Fig. 1) showed excellent in vitro potency and attractive oral bioavailability and efficacy in TB mouse models (12). Interestingly, resistance against this new drug candidate emerged at extremely low frequency and was associated with strong fitness costs (12). Genome analyses revealed that resistance was associated with amplification of large chromosomal segments, all containing the dnaN gene, suggesting dnaN overexpression as a mechanism of resistance (12). Indeed, binding studies and costructural analyses showed that griselimycins target mycobacterial DnaN (12).
DnaN encodes the DNA sliding clamp, also referred to as DNA polymerase III b subunit. This DNA sliding clamp is crucial for bacterial DNA replication and repair, acting as a proteinprotein interaction (PPI) hub. The protein surrounds double-stranded DNA and functions to recruit a diverse range of accessory proteins involved in DNA metabolism (16)(17)(18). DnaN  protein partners interact with a specific hydrophobic cleft on the DnaN clamp. Griselimycins bind to the same cleft, disrupting DnaN PPIs, as shown by elegant biochemical analyses (12). On-target activity was recently confirmed by de Wet and colleagues in intact mycobacteria (19). By combining inducible CRISPR interference and image-based analyses of morphological features in mycobacteria, the authors demonstrated that griselimycin copied the phenotype of a dnaN knockdown (19). Fluorescence microscopy analyses further demonstrated that griselimycins cause replisome instability and affect the structure of the nucleoid in vivo (20). Thus, the peptide antibiotic griselimycin corrupts DnaN-dependent machines involved in genome copying and maintenance by acting as a PPI inhibitor. Interestingly, CGM not only was potent in vitro against Mycobacterium tuberculosis but also was active against the nonpathogenic mycobacterial model organism Mycobacterium smegmatis (12). To determine whether CGM retained activity against the opportunistic pathogen M. abscessus, we measured its MIC against reference strains and clinical isolates of the three M. abscessus complex subspecies, using CGM from Evotec's compound archive (12). Dose-response curves were established using the broth dilution method in Middlebrook 7H9 medium (BD Difco) and optical density at 600 nm (OD 600 ) as the readout for growth (21). CGM exhibited uniform submicromolar growth-inhibitory activity against all M. abscessus strains tested ( Table 1), suggesting that CGM is broadly active against the M. abscessus complex.
To determine whether CGM retained bactericidal activity as observed against M. tuberculosis (12), dose-response time-kill experiments were carried out with the type strain M. abscessus ATCC 19977 (21). Treatment with CGM at the MIC (0.5 mM) resulted in 10-fold and .1,000fold reductions in CFU after 1 and 3 days, respectively, indicating pronounced time-dependent bactericidal activity ( Fig. 2A). Time kill experiments were also carried out for M. abscessus K21, the strain we employ in our mouse infection studies (see below). Interestingly, the bactericidal activity of CGM against M. abscessus K21 was lower than that against M. abscessus ATCC 19977. Despite showing similar MICs against both strains (;0.5 mM) (Table 1), higher concentrations of CGM were required to achieve comparable reduction of CFU in M. abscessus K21 cultures (Fig. 2B). The reason for the apparent strain-dependent bactericidal activity of CGM remains to be determined. Figure 2C and D show the results of the time-kill experiments for the mostly bacteriostatic clarithromycin as control. Consistent with previous results (22), treatment with the macrolide did not result in significant reduction of CFU.
To assess the in vivo efficacy of CGM, we infected 8-week-old female NOD.CB17-Prkdc scid /NCrCrl (NOD SCID) mice (Charles River Laboratories) by intranasal delivery of 10 6 CFU M. abscessus K21 as described previously (23). In this immunodeficient mouse model, the K21 strain produces a sustained infection resulting in a largely constant bacterial lung burden over time, thus allowing the effects of drugs to be evaluated (23). Drugs or the vehicle control was administered once daily for 10 consecutive days by oral gavage, starting 1 day postinfection. CGM was formulated in Cremophor RH 40-Capryol 90-Miglyol 812 N (10/20/70 [wt/wt/wt]) and administered at 250 mg/kg of body weight. Clarithromycin, formulated in 0.4% methylcellulose-sterile water, was used as a positive control at the human-equivalent dose of 250 mg/kg. All mice were euthanized 24 h after the last dose, and bacterial load in the lungs and spleen was determined by plating serial dilutions of the organ homogenates onto Middlebrook agar. All experiments involving live animals were approved by the Institutional Animal Care and Use Committee of the Center for Discovery and Innovation, Hackensack Meridian Health. As expected, treatment with the vehicle did not affect the bacterial lung burden (Fig. 3A). Compared to the vehicle control, treatment with 250 mg/kg CGM reduced lung CFU 10-fold and thus more than the positive-control clarithromycin at 250 mg/kg (Fig. 3A). CFU reduction in the spleen followed a similar pattern (Fig. 3B). Thus, CGM is efficacious in a mouse model of M. abscessus infection. In conclusion, we show that the cyclohexyl analog of griselimycin, CGM, is broadly active against the M. abscessus complex in vitro. The advanced anti-TB lead compound displayed bactericidal activity in vitro and reduced the bacterial lung burden in a mouse model of M. abscessus infection. This work adds a new advanced lead compound to the preclinical M. abscessus drug discovery pipeline and suggests that the new anti-TB drug candidate could be explored for the treatment of M. abscessus lung disease. The demonstration that yet another TB active displays anti-M. abscessus activity supports the paradigm of exploiting chemical matter generated for TB drug discovery to accelerate de novo drug discovery for M. abscessus.

ACKNOWLEDGMENTS
We are grateful to Wei Chang Huang (Taichung Veterans General Hospital, Taichung