Characterization of Two Novel Inhibitors of the Mycobacterium tuberculosis Cytochrome bc₁ Complex

ABSTRACT

Drug-resistant tuberculosis is a global health care threat calling for novel effective treatment options. Here, we report on two novel cytochrome bc₁ inhibitors (MJ-22 and B6) targeting the Mycobacterium tuberculosis respiratory chain with excellent intracellular activities in human macrophages. Both hit compounds revealed very low mutation frequencies and distinct cross-resistance patterns with other advanced cytochrome bc₁ inhibitors.

TEXT

After a constant decline of new tuberculosis (TB) cases in the past decades, 2021 marks the first year with an increase of 0.5 million tuberculosis cases compared to the previous year. Among these 10.6 million cases there were an estimated 450,000 cases of multidrug-resistant TB (MDR-TB), which no longer respond to the standard antibiotic regimen (1). This illustrates that there is an urgent need for lead compounds with novel modes of action active against drug-resistant Mycobacterium tuberculosis. The past decade revealed that the M. tuberculosis respiratory chain is a highly susceptible target for synthetic small molecules (2–4). With bedaquiline, inhibiting the ATP synthase, and Q203, targeting the cytochrome bc₁ complex, two respiratory chain inhibitors have been approved by regulatory authorities or are currently being investigated in phase 2 clinical trials (5, 6). Screening of 60,000 small molecules using a phenotypic resazurin microtiter assay (REMA) (7) at single concentrations (10 μM) identified 425 hit compounds (0.71% hit rate) with antituberculous activity (8, 9). A subsequent secondary screening based on a fibroblast-associated infection model (10) allowed for exclusion of cytotoxic compounds and indicated potential intracellular antibiotic activity. After confirmation of activity in 2-fold serial dilutions, this dual screening yielded 11 compounds (0.02% hit rate) with potential intracellular antibiotic activity. Out of these hit compounds we selected two substances, namely, B6, a quinoline-amine, and MJ-22, a thiazole, for further investigations (Fig. 1A). Both compounds showed good antituberculous activity with minimal inhibitory concentrations (MICs) of 1.12 μM (B6) and 3.69 μM (MJ-22) in 7H9 broth using the REMA (Fig. 1B) (7). Intracellular activity was determined using enhanced green fluorescent protein (eGFP)-expressing THP-1-derived macrophages which were infected with dsred2-expressing M. tuberculosis. To confirm that the red fluorescent signal obtained in this assay correlates with the intracellular mycobacterial load, we first generated a standard curve, correlating fluorescence with the number of bacteria in macrophages (see Fig. S3 and the methods section in the supplemental material). This correlation analysis between CFU extracted from infected macrophages and the fluorescent signal provided an R² value of 0.81, indicating that the measured dsred2 signal is a good surrogate for the number of intracellular bacteria. Using this assay, we treated intracellular bacteria with the two compounds, which led to a dose-dependent reduction of intracellular growth of M. tuberculosis with a 50% inhibitory concentration (IC₅₀) of 1.78 μM for B6 and 0.55 μM for MJ-22 (Fig. 1C and D). MJ-22 did not lead to a full suppression of the dsred2 signal compared to B6. The reason for this is not fully understood. Nevertheless, a clear concentration-dependent effect could be observed for both compounds. Further investigation of the infected macrophages also revealed strongly increased host cell survival following treatment, which is an additional surrogate for excellent intracellular activity of antimycobacterial drugs (Fig. 1E). Using an additional, highly sensitive assay which quantifies survival of M. tuberculosis-infected MRC-5 human lung fibroblasts (10), we determined 50% effective concentrations (EC₅₀) of 0.78 μM (B6) and 1.36 μM (MJ-22) (Fig. S1). No cytotoxic effects were observed for MJ-22 in exposed THP-1 macrophages at concentrations up to 200 μM after 72 h, whereas B6 showed cytotoxicity only at very high concentrations (IC₅₀ of 89.37 μM), providing a selectivity index (SI) of 79.8 (Fig. S2). To elucidate the modes of action, we performed a series of secondary assays including quantification of intracellular ATP pools, which revealed that exposure to either of the compounds strongly reduced ATP levels in M. tuberculosis, suggesting the respiratory chain as a potential molecular target (Fig. 1F). The known cytochrome bc₁ complex inhibitors Q203 and lansoprazole sulfide (LPZS) were used as positive controls (11, 12). Based on these findings, we focused on potential effects on components of the respiratory chain and could show that treatment of M. tuberculosis with either compound led to significant upregulation of cydB (Rv1622c) in reverse transcriptase quantitative PCR (RT-qPCR) experiments (Fig. 1G). The gene cydB codes for a component of the alternative terminal oxidase cytochrome bd. Upregulation of cytochrome bd represents a compensatory mechanism of M. tuberculosis which partly restores activity of the respiratory chain following chemical inhibition of cytochrome bc₁, indirectly showing that cytochrome bc₁ is the putative target of B6 and MJ-22 (13). In contrast, deletion of cytochrome bd facilitates the generation and selection of spontaneous cytochrome bc₁ mutants resistant to the respective inhibitors (13). By plating bacteria on agar containing the compounds at concentrations five times above the MIC, we were able to generate a single resistant clone for B6 and MJ-22, respectively. We were not able to raise additional clones despite several attempts. Based on the number of bacteria plated for these experiments, we calculated the frequency of resistance for both compounds (B6, 2.17 × 10⁻¹⁰; MJ-22, 3.33 × 10⁻⁹). For Q203 we calculated a frequency of 5 × 10⁻⁹ (own data versus 6.25 × 10⁻⁹ described by Lee et al. [14]). Thus, at least for B6, the frequency is considerably lower than for Q203 and other cytochrome bc₁ inhibitors (14–17). Whole-genome sequencing of the resistant clones confirmed mutations exclusively in the cytochrome bc₁ complex b-subunit qcrB (Rv2196) at codon 182 (S182T) for B6 and at codon 342 (M342V) for MJ-22. This provides further evidence that cytochrome bc₁ is the molecular target of B6 and MJ-22. Testing both compounds for cross-resistance to the most frequent resistance-conveying mutations in other cytochrome bc₁ inhibitors such as T313A (12, 18) and A178V (15) resulted in a distinct pattern. While both B6 and MJ-22 showed an increase of MIC in QcrB mutants with S182T/P, A178V, and M342V amino acid changes or the QcrA (Rv2195) L356V/W amino acid change, no resistant phenotype was observed in the QcrB T313A mutant, which is a key driver of resistance for the most advanced cytochrome bc₁ inhibitor, Q203 (Table 1 and Fig. S4) (12). Interestingly, the T313A mutant strain showed antibiotic hypersensitivity to B6, with a 36-fold increase in susceptibility and an MIC in the low nanomolar range (Fig. 1H). The mechanism for this interesting phenotype requires further investigations.

FIG 1.

Small molecules show antituberculous activity, host cell protection, and inhibition of the respiratory chain of Mycobacterium tuberculosis. (A and B) The two compounds (A) B6 (left) and MJ22 (right) show high efficacy (B) in 7H9 broth with MIC₅₀s of 1.62 μM and 3.69 μM, respectively. (C) Representative microscopy pictures 48 h after infection with dsred2-expressing M. tuberculosis (yellow) of eGFP-expressing THP-1-derived macrophages (blue) (Rif, rifampicin; DMSO, dimethyl sulfoxide). (D) Intracellular growth of dsred2-expressing M. tuberculosis 48 h postinfection (multiplicity of infection of 2) after exposure to different concentrations of B6 (left) or MJ-22 (right). Data are normalized to 0.1% DMSO (100%) and 10 μM rifampicin (0%). (E) Cell count of eGFP-expressing THP-1-derived macrophages 48 h postinfection (multiplicity of infection of 2). Data are normalized to rifampicin (Rif; survival = 100%) or 0.1% DMSO (DMSO; host cell death = 0%). M. tuberculosis dsred2 signal and THP-1 cell count were determined using ImageJ. (F) Relative intracellular ATP levels after 24 h of exposure to the respective compound normalized to 0.1% DMSO. (G) Relative expression of the cytochrome bd subunit II cydB (Rv1622c) after 4 h of exposure and normalized to 0.1% DMSO. (H) Antituberculous activity of B6 against the ΔcydAB parental strain (blue) and ΔcydAB mutants containing the QcrB mutation T313A (orange) or S182T (brown). Data are shown as mean and standard error of the mean. Box plot whiskers represent minimal and maximum values. Statistical significances are compared to DMSO based on one-way analysis of variance with a post hoc pairwise comparison and Bonferroni correction (**, P < 0.01; ***, P < 0.001). Scale bars shown in microscopy images represent 200 μm.

TABLE 1.

Cross-resistance patterns of various resistance-facilitating mutations^a

Compound	M. tuberculosis H37Rv ΔcydAB (MIC50)
	Parental	QcrB					QcrA
		S182T	S182P	A178V	T313A	M342V	L356V	L356W
Q203 (nM)	1.27 (0.98–1.47)	1.57 (1.04–2.37)	5.84 (4.55–6.19)	0.86 (0.75–1.01)	>100 (NA)	13.01 (12.17–16.72)	13.05 (11.68–16.58)	>100 (NA)
MJ-22 (μM)	4.89 (4.48–5.28)	>50 (NA)	>50 (NA)	>50 (NA)	7.43 (6.57–9.74)	>50 (NA)	>50 (NA)	>50 (NA)
B6 (μM)	1.12 (0.86–1.22)	8.43 (6.66–9.51)	8.90 (7.00–9.84)	5.38 (3.76–8.09)	0.03 (0.03–0.03)	8.43 (6.90–9.49)	7.94 (6.66–9.40)	6.14 (5.34–8.46)

^aThe MIC₅₀s of Q203, B6, and MJ-22 against the M. tuberculosis ΔcydAB parental strain and mutants harboring mutations in the QcrB (Rv2196) (S182T, S182P, A178V, T313A, and M342V) and QcrA (Rv2195) (L356V and L356W) loci were tested in resazurin microtiter assays. The lower and upper confidence levels of the MIC₅₀ are indicated in parentheses. Dose-response curves of each compound to the respective mutant are displayed in Fig. S4 in the supplemental material. NA, not applicable. Boldface highlights most common mutation in Q203 resistant mutants.

To conclude, with B6 and MJ-22 we identified two promising antituberculous hit compounds targeting the M. tuberculosis respiratory chain. Both substances showed good intracellular activity and selectivity indices regarding cytotoxicity. Although compound B6 was previously described in an independent screening focusing on ATP homeostasis inhibitors (19), with cytochrome bc₁, we were now able to fully elucidate the target of this hit compound. Lack of cross-resistance or even hypersensitivity in M. tuberculosis strains with reduced efficacy of the frontrunner drug Q203 or other cytochrome bc₁ inhibitors may allow for extended clinical application of B6 and MJ-22 derivatives in pretreated individuals. Combining two cytochrome bc₁ inhibitors with well-defined and distinct resistance mechanisms could be another possibility to combat the emerging problem of multidrug resistance.