Dual inhibition of the terminal oxidases eradicates antibiotic-tolerant Mycobacterium tuberculosis

Abstract

The approval of bedaquiline has placed energy metabolism in the limelight as an attractive target space for tuberculosis antibiotic development. While bedaquiline inhibits the mycobacterial F₁ F₀ ATP synthase, small molecules targeting other components of the oxidative phosphorylation pathway have been identified. Of particular interest is Telacebec (Q203), a phase 2 drug candidate inhibitor of the cytochrome bcc:aa₃ terminal oxidase. A functional redundancy between the cytochrome bcc:aa₃ and the cytochrome bd oxidase protects M. tuberculosis from Q203-induced death, highlighting the attractiveness of the bd-type terminal oxidase for drug development. Here, we employed a facile whole-cell screen approach to identify the cytochrome bd inhibitor ND-011992. Although ND-011992 is ineffective on its own, it inhibits respiration and ATP homeostasis in combination with Q203. The drug combination was bactericidal against replicating and antibiotic-tolerant, non-replicating mycobacteria, and increased efficacy relative to that of a single drug in a mouse model. These findings suggest that a cytochrome bd oxidase inhibitor will add value to a drug combination targeting oxidative phosphorylation for tuberculosis treatment.

Keywords: Q203; antibiotic-tolerance; cytochrome bcc-aa3; cytochrome bd oxidase; oxidative phosphorylation.

Figure 1. Identification of the putative Cyt‐bd inhibitor ND‐011992

1. Molecular structure of ND‐011992.

2. Effect of ND‐011992 on the intracellular ATP level in M. bovis BCG. The bacteria were treated with DMSO (blue triangle), 100 nM Q203 alone (green triangle), 20 μM ND‐011992 alone (brown square), and a dose range of ND‐011992 in the presence of Q203 (red squares) for 15 h before quantification of intracellular ATP levels.

3. Oxygen consumption assay in M. bovis BCG using methylene blue as an oxygen sensor.

4. Oxygen consumption assay in M. bovis BCG using the Seahorse XFe96 Extracellular flux analyzer. 12 μM ND‐011992 was injected alone (green triangles), 100 nM Q203 (blue circles), or in combination with Q203 (red squares). For each condition, OCR readings were normalized to the last basal OCR reading before drug injection. OCR: oxygen consumption rate. The dotted lines indicate the timepoints of drug injection.

5. Dose‐dependent inhibition of M. bovis BCG OCR by ND‐011992 (in the presence of 100 nM Q203) measured on a Seahorse XFe96 analyzer platform. For each condition, OCR readings were normalized to the last basal OCR reading before drug injection.

6. Effect of ND‐011992 on the OCR of M. smegmatis IMVs using the Oroboros O₂k fluorespirometer. IMV OCR from the parental strain (green triangles), ∆cydAB knockout (blue circles), and ∆cydAB complemented with M. tuberculosis CydABDC⁺ (red squares) energized with NADH were determined. Q203 was used at 1 μM. 100% OCR was defined as the OCR of the untreated samples for each strain.

Data information: Data are expressed as the mean ± SD for each condition of a representative experiment. Experiments in (B) and (E) were performed with 2 technical replicates and independently repeated at least twice. In (D) and (F), each experiment was performed with 3 technical replicates and independently repeated at least once.

Figure EV1. ND‐011992 synergizes with Q203 to inhibit ATP production and growth in mycobacteria

1. Effect of ND‐011992 on the intracellular ATP level in M. tuberculosis H37Rv. The bacteria were treated with DMSO (blue square), 100 nM Q203 (green triangle), 25 µM ND‐011992 (brown triangle), and varying concentrations of ND‐011992 in the presence of 100 nM Q203 (red circles) for 15 h.

2. Isobologram analysis reflecting the interactions between ND‐011992 and Q203 in an intracellular ATP depletion checkerboard assay in M. bovis BCG.

3. Isobologram analysis reflecting the interactions between ND‐011992 and Q203 in an intracellular ATP depletion checkerboard assay in M. tuberculosis H37Rv.

4. Effect of ND‐011992 on the dose‐dependent growth inhibition curve of Q203 in M. tuberculosis H37Rv. Blue triangles: Q203 alone; purple triangles: with 3.13 μM ND‐011992; red squares: with 12.5 μM ND‐011992. Data are expressed as the mean ± SD of triplicates for each condition.

Figure EV2. ND‐011992‐Q203 combination inhibits oxygen respiration in mycobacteria

1. Oxygen consumption assay in M. tuberculosis H37Rv using methylene blue as an oxygen sensor.

2. Oxygen consumption assay in M. tuberculosis H37Rv using the MitoXpress Xtra® probe. H37Rv was incubated with the oxygen probe in the presence of 1% DMSO (red circles), 100 nM Q203 (green squares), 25 μM ND‐011992 (blue triangles), or 25 μM ND‐011992 + 100 nM Q203 (pink diamonds).

3. Oxygen consumption rate of M. tuberculosis H37Rv derived from the MitoXpress Xtra® assay (panel B). **P = 0.0036, unpaired Student’s t‐test; two‐tailed; n = 3.

4. Oxygen consumption assay in M. bovis BCG using the MitoXpress Xtra® probe. M. bovis BCG was incubated with the oxygen probe in the presence of 1% DMSO (black circles), 100 nM Q203 (blue squares), 25 μM ND‐011992 (dark green triangles), or 25 μM ND‐011992 + 100 nM Q203 (light green triangles).

5. Oxygen consumption rate of M. tuberculosis H37Rv derived from the MitoXpress Xtra® assay (panel D). **P = 0.00000042, unpaired Student’s t‐test; two‐tailed; n = 3.

6. Effect of ND‐011992 on the ATP synthesis of M. tuberculosis H37Rv mc²6230 inverted‐membrane vesicles (IMV) using the Roche Bioluminescence Assay Kit CL II. Q203 was used at 100 nM, and ND‐011992 was used at 40 μM. The figure shows averages and standard deviations of triplicate repeats with the blank subtracted from all conditions. P‐value between drug‐free and ND‐011992: 0.1159; **P = 0.0020, unpaired Student’s t‐test; two‐tailed; n = 3.

Data information: Data are expressed as the mean ± SD of triplicates for each condition.

Figure 2. ND‐011992 inhibits the mycobacterial Cyt‐bd

1. Differential gene expression analysis of H37Rv treated with Q203, ND‐011992, or combination (H37Rv‐Combo), and ∆cydAB treated with Q203 compared with untreated controls. The log₂‐fold differences in gene expression of various conditions relative to the untreated control are indicated using a sliding scale where higher expression is reflected in red and lower expression is reflected in blue, with a midpoint signifying no difference in white.

2. Effect of cydABDC overexpression on the oxygen consumption rate (OCR) and intracellular ATP levels. In M. bovis BCG treated with ND‐011992 in the presence of 100 nM Q203. IC₅₀: inhibitory concentration 50%.

3. Absence of the Cyt‐bcc:aa₃ sensitized H37Rv to ND‐011992. M. tuberculosis H37Rv wild type (blue circles), ∆ctaE‐qcrCAB (red squares), and complemented strain (green triangles) were treated with ND‐011992. Growth inhibitory potency (MIC₅₀) was recorded after 2–3 weeks of incubation.

4. The reduced minus oxidized difference spectrum of M. smegmatis ΔqcrCAB inverted‐membrane vesicles (IMVs). The green line represents the spectrum without treatment, while the blue line represents the spectrum of the IMVs in the presence of 10 μM ND‐011992.

Data information: Data are expressed as the mean ± SD of triplicates for each condition of a representative experiment. Each experiment was performed with 3 technical replicates and independently repeated at least once. Source data are available online for this figure.

Figure EV3. ND‐011992 does not affect electron transfer within the cytochrome bcc:aa₃ supercomplex

1. A

   NADH‐driven electron transfer within plasma membrane vesicles (IMVs) of the wild‐type (gray) and ΔqcrCAB M. smegmatis (green). The difference spectrum of IMVs from WT M. smegmatis after addition of the electron donor NADH shows the major absorbance peaks for heme a, b, and c of the cytochrome bcc‐aa₃ supercomplex. The characteristic peaks of Cyt‐bd are displayed after reduction by 2 mM NADH in the ΔqcrCAB background.

2. B, C

   Effect of Q203 and ND‐011992 on NADH‐driven electron transfer in wild‐type M. smegmatis IMVs. (B) The NADH‐based difference spectra in the absence (black) and presence of 7.5 µM Q203 (red) reflect the effect of Q203 on the reduction of hemes in the cytochrome bcc‐aa₃. (C) The NADH‐based difference spectra in the absence (black) and the presence of 10 µM ND‐011992 (orange).

Figure 3. ND‐011992 has a low spontaneous resistance mutation frequency and is active against drug‐resistant M. tuberculosis clinical isolates

1. Combined potency of ND‐011992 + Q203 against the drug‐resistant clinical isolate M. tuberculosis 123‐20‐0047.

2. Potency of the combination ND‐011992 + Q203 against various clinical isolates in an agar plate MIC assay. A dose range of ND‐011992 was tested against the clinical isolates in the presence of Q203. Q203 was used at 100 nM against the strains H37Rv, N1274, N1283, 123‐20‐0015, 123‐20‐0091, 123‐20‐0041, 123‐20‐0047, and at 20 nM against the stains N0072, N0157, N0145, N0052, and N0136. “S” indicates pan‐susceptibility to anti‐TB drugs; “MDR”: multi‐drug resistant; “XDR”: extensively drug‐resistant.

3. Growth inhibition assay in M. bovis BCG. 10⁶ bacteria were plated onto 7H10 agar supplemented with DMSO, 25 nM Q203, 3 µM ND‐011992, or 25 nM Q203 + 3 µM ND‐011992.

4. Fluctuation analysis in M. tuberculosis H37Rv. M. tuberculosis ΔcydAB was plated on 7H10 containing 100 nM Q203 (red). Parental H37Rv was plated on 7H10 with 100 nM Q203 and 6 µM ND‐011992 or 2 µg/ml rifampicin (blue). The corresponding values of the median frequency of resistance are indicated in the graph.

Figure 4. The combination ND‐011992/Q203 is bactericidal against replicating and non‐replicating M. tuberculosis H37Rv and shows potency in vivo

1. Bactericidal activity of ND‐011992 and ND‐011992‐Q203 combination in replicating M. tuberculosis H37Rv. Bedaquiline (BDQ) was used at 500 nM, and Q203 at 100 nM. Inoc.: starting inoculum; DMSO: vehicle control. Bacteria viability was determined by enumerating colony‐forming units (CFU) on nutrient‐agar plates after 15 days of incubation. Dotted line represents 90% reduction in CFU/ml compared with the initial inoculum at time 0. *P < 0.05; **P < 0.01 unpaired Student’s t‐test, one‐tailed; n = 3; comparing Q203 alone vs Q203 + ND‐011992. Exact P‐values (from left to right): 0.0346, 0.0143, 0.00016, 0.000052.

2. Bactericidal activity of ND‐011992 and ND‐011992‐Q203 combination in nutrient‐starved M. tuberculosis H37Rv. Bedaquiline (BDQ) was used at 500 nM, isoniazid was used at 1 μM, Q203 was used at 100 nM, and ND‐011992 was used at 3, 6, 12, and 30 μM. Bacteria viability was determined by enumerating colony‐forming units (CFU) after 15 days of incubation. Dotted line represents 90% reduction in CFU/mL compared with the initial inoculum at time 0. **P < 0.01 unpaired Student’s t‐test, one‐tailed; n = 3; comparing Q203 alone vs Q203 + ND‐011992. Exact P‐values (from left to right): 0.0098, 0.000027, 0.000067.

3. Bactericidal activity of ND‐011992 and ND‐011992‐Q203 combination in M. tuberculosis H37Rv under hypoxia. BDQ was used at 500 nM, isoniazid at 1 μM, metronidazole (MTZ) at 200 μM, Q203 at 100 nM, and ND‐011992 at 3, 6, 12, and 30 μM. Bacteria viability was determined by enumerating colony‐forming units (CFU) after 15 days of incubation. Dotted line represents 90% reduction in CFU/mL compared with the initial inoculum at time 0.**P < 0.01 unpaired Student’s t‐test, one‐tailed; n = 3; comparing Q203 alone vs. Q203 + ND‐011992. Exact P‐values (from left to right): 0.0092, 0.00010, 0.00079.

4. Efficacy of ND‐011992‐Q203 combination treatment in a mouse model of acute tuberculosis. Bacterial load (CFU) was enumerated in the lungs of mice before treatment, and after 5 days of treatment with either the vehicle control (brown squares), Q203 at 5 mg/kg (red squares), ND‐011992 at 25 mg/kg (green triangles), or ND‐011992 at 25 mg/kg with Q203 at 5 mg/kg (orange diamonds). **P = 0.0006 unpaired Student’s t‐test, two‐tailed; n = 4; comparing Q203 alone vs. Q203 + ND‐011992.

Data information: Data are expressed as the mean ± SD of triplicates for each data point. Experiments in (A), (B), and (C) were performed with 3 technical replicates and independently repeated at least once. In vivo experiment in (D) was performed with 4 replicates for all treatment groups, and 3 replicates to elucidate bacterial load before initiation of treatment. The experiment was repeated once.

Figure EV4. The drug combination ND‐011992‐Q203 is bactericidal against Mycobacterium bovis BCG

1. Bactericidal activity of ND‐011992 and ND‐011992‐Q203 combination in replicating M. bovis BCG. Q203 was used at 100 nM, and ND‐011992 used at 3, 6, 12, and 30 μM alone or in combination with Q203. Dotted line represents 90% reduction in CFU/ml compared with the initial inoculum at time 0. Exact P‐values (from left to right): 0.0002, 0.0002, 0.0003, 0.00002.

2. Bactericidal activity of ND‐011992 and ND‐011992‐Q203 combination in M. bovis BCG under oxygen‐starved condition. Metronidazole (MTZ) was used at 200 μM, Q203 was used at 100 nM, and ND‐011992 was used at 3, 6, 12, and 30 μM. Bacterial viability was determined via CFU enumeration after 10 days of treatment. Dotted line represents 99% reduction in CFU/mL compared with the initial inoculum at time 0. Exact P‐values (from left to right): 0.0024, 0.0075, 0.0029, 0.0000003.

Data information: **P < 0.01, unpaired Student’s t‐test, one‐tailed; n = 3; comparing Q203 alone vs Q203 + ND‐011992. Data are expressed as the mean ± SD of triplicates for each condition.