Towards a new combination therapy for tuberculosis with next generation benzothiazinones

Abstract

The benzothiazinone lead compound, BTZ043, kills Mycobacterium tuberculosis by inhibiting the essential flavo-enzyme DprE1, decaprenylphosphoryl-beta-D-ribose 2-epimerase. Here, we synthesized a new series of piperazine-containing benzothiazinones (PBTZ) and show that, like BTZ043, the preclinical candidate PBTZ169 binds covalently to DprE1. The crystal structure of the DprE1-PBTZ169 complex reveals formation of a semimercaptal adduct with Cys387 in the active site and explains the irreversible inactivation of the enzyme. Compared to BTZ043, PBTZ169 has improved potency, safety and efficacy in zebrafish and mouse models of tuberculosis (TB). When combined with other TB drugs, PBTZ169 showed additive activity against M. tuberculosis in vitro except with bedaquiline (BDQ) where synergy was observed. A new regimen comprising PBTZ169, BDQ and pyrazinamide was found to be more efficacious than the standard three drug treatment in a murine model of chronic disease. PBTZ169 is thus an attractive drug candidate to treat TB in humans.

Figure 1

Chemical structures. Structure of BTZ043 and five compounds of the PBTZ family selected for in vivo studies, PBTZ169, PBTZ139, PBTZ134, PBTZ137 and PBTZ139.

Figure 2

NfnB-mediated reduction of BTZ043 and PBTZ169. Compounds (50 μM) were incubated with NfnB (6 μM) as described in Materials and Methods. Samples (10 μL) were taken at different times during the reaction and directly analyzed by LC-MS (ESI+). Results were analyzed by extracting the ion counts for the species of interest for each compound, namely the protonated forms of the original compound, and its nitroso, hydroxylamino and amino forms, as well as the azoxy adducts.

Figure 3

Crystal structure of the M. tuberculosis DprE1-PBTZ169 complex.

1. Cartoon representation of the M. tuberculosis DprE1 structure (grey) in complex with PBTZ169 adduct (green sticks), and its FAD cofactor (yellow sticks), superposed on the M. smegmatis DprE1 structure (blue) in complex with BTZ043 and FAD. The location of the disordered regions in the M. tuberculosis DprE1 are marked as black dashed lines. Residue numbers are indicated in black for the M. tuberculosis structure and in blue for the M. smegmatis protein.

2. Close-up view of the M. tuberculosis DprE1 active site showing the residues in close contact with PBTZ169. The binding pocket of the CF₃ group of PBTZ169 is shown as a surface representation.

Figure 4

Treatment of M. marinum infected zebrafish embryos with BTZ. A One hour post infection (1 hpi) zebrafish embryos that had been infected with 50-200 CFU of fluorescent M. marinum were treated for 5 days with BTZ043 or PBTZ169 at the concentrations indicated. Bacterial burden was assessed by fluorescence microscopy using a Leica MZ16FA microscope and customized software to quantify infection levels. Graphs represent data from 2 - 8 independent experiments: ns, not significant; **P < 0.001; ***P < 0.0001; one-way ANOVA, Bonferroni's multiple comparison test. The M strain was used for infection. B–F Images of infected zebrafish embryos treated with PBTZ169 at 0 (WT), 5, 25 and 50 nM, or BTZ043 at 50 nM. The arrows indicate clusters of mycobacteria that were not present in fish treated with (P)BTZ at concentrations >25 nM. Note that the red fluorescence on the yolk/gut of the embryos is background staining.

Figure 5

Effect of PBTZ169 treatment on infected zebrafish embryos. Zebrafish embryos were infected with either M. marinum strain M (A, B) or E11 (C, D). Treatment with PBTZ169 (25 μM) started at 0, 1 or 2 days post-infection (dpi) up to 5 dpi, giving exposure times of 5, 4 and 3 days, respectively. The bacterial burden was assessed by fluorescence microscopy and customized software to quantify the number of fluorescent pixels (A and C), and by enumeration of the CFU in lysed embryos (B and D), immediately after imaging was performed. In B and D, each data point represents the mean number of CFU per embryo of 3 embryo extracts plated, the bars indicate means and s.e.m. after log transformation. The graphs represent data from 2 to 4 independent experiments. ***P < 0.0001, one-way ANOVA, Bonferroni's multiple comparison test. A,B Zebrafish embryos infected with M. marinum strain M. C,D Zebrafish embryos infected with M. marinum strain E11.

Figure 6

Effect of BTZ043 on M. marinum infected zebrafish embryo development. A–D M. marinum infected zebrafish embryos were treated with BTZ043 for 5 days at the concentrations indicated and photographed using a Leica DFC420C camera. The M strain was used for infection. E–F Images of the notochord from untreated and treated zebrafish, arrows indicate developmental abnormalities. G Analysis of zebrafish embryos displaying developmental abnormalities treated with BTZ043 (BTZ) or PBTZ169 (PBTZ) at the concentrations indicated.

Figure 7

Efficacy studies in vivo.

1. Efficacy of 5 PBTZ candidates in a mouse model of chronic TB compared with INH (25 mg/kg), BTZ043 and untreated controls (NT). All BTZs were administered at 50 mg/kg of body weight per day. Red and black columns correspond to the bacterial burden in the lungs and spleens, respectively, at day 0 (D0) when treatment initiated, or day 28 (D28) when treatment finished. Bars represent the mean ± s.d. of CFUs from 5 mice per group. Significance in difference relative to BTZ043 were calculated using Student's t-test. *P < 0.05; **P < 0.005.

2. Dose escalation study of PBTZ169 in the same model. 5 mice per group were treated with various drugs at the doses indicated (mg/kg). Colors, bars and statistics are as in (A).

Figure 8

Synergistic interaction in vitro between BDQ and PBTZ169. The effect of serial dilutions of PBTZ169 and BDQ in combination on the viability of M. tuberculosis H37Rv was assessed after 7 days exposure. Bacterial dilutions were plated on solid medium and incubated for 1 month before counting the number of CFU. Drug concentrations are in ng/ml.

Figure 9

Efficacy of PBTZ169-based combination treatment in mice. The bactericidal effects of PBTZ, BDQ (both 25 mg/kg) and PZA (150 mg/kg) alone and in combination were compared in the lung (A) and in the spleen (B) with a standard regimen of RIF, INH and PZA (RHZ) and untreated controls (NT) in the murine model of chronic TB. The combination of PBTZ169 and BDQ plus PZA was significantly more effective than the standard treatment in reducing the number of CFU in both organs after 1 and 2 months of treatment (Supplementary Table 6). The lung CFU count of PBTZ-BDQ-PZA (*) was below the limit of detection used after 4 weeks of treatment (<200 CFU).