Mode of Action of Clofazimine and Combination Therapy with Benzothiazinones against Mycobacterium tuberculosis

Abstract

Clofazimine (CZM) is an antileprosy drug that was recently repurposed for treatment of multidrug-resistant tuberculosis. In Mycobacterium tuberculosis, CZM appears to act as a prodrug, which is reduced by NADH dehydrogenase (NDH-2), to release reactive oxygen species upon reoxidation by O₂. CZM presumably competes with menaquinone (MK-4), a key cofactor in the mycobacterial electron transfer chain, for its reduction by NDH-2. We studied the effect of MK-4 supplementation on the activity of CZM against M. tuberculosis and found direct competition between CZM and MK-4 for the cidal effect of CZM, against nonreplicating and actively growing bacteria, as MK-4 supplementation blocked the drug's activity against nonreplicating bacteria. We demonstrated that CZM, like bedaquiline, is synergistic in vitro with benzothiazinones such as 2-piperazino-benzothiazinone 169 (PBTZ169), and this synergy also occurs against nonreplicating bacteria. The synergy between CZM and PBTZ169 was lost in an MK-4-rich medium, indicating that MK-4 is the probable link between their activities. The efficacy of the dual combination of CZM and PBTZ169 was tested in vivo, where a great reduction in bacterial load was obtained in a murine model of chronic tuberculosis. Taken together, these data confirm the potential of CZM in association with PBTZ169 as the basis for a new regimen against drug-resistant strains of M. tuberculosis.

INTRODUCTION

With approximately 9 million incident cases of tuberculosis (TB) worldwide and around 1.5 million deaths in 2012, Mycobacterium tuberculosis infection is one of the most important causes of death from a single infectious agent (1). The spread of multidrug-resistant TB (MDR-TB), namely, with resistance to isoniazid and rifampin, poses additional challenges to treatment with currently available anti-TB drugs. The situation is exacerbated by the increasing emergence of extensively drug-resistant (XDR) strains of M. tuberculosis, which cause diseases essentially untreatable with existing compounds. It is nowadays widely acknowledged that we need to develop new antibiotic combinations for TB and that these new regimens should be tested together at the preclinical stage, rather than testing a series of single drugs separately, in order to fill the TB drug development pipeline more efficiently (2–4).

Some of the compounds in advanced clinical trials for TB are molecules that were originally used to treat other infectious diseases and have been repurposed for TB. Among the repurposed molecules, clofazimine (CZM), a riminophenazine originally developed as a drug to treat TB but overlooked for decades, has been used as a standard component of the treatment of leprosy for 50 years. It was recently repurposed for managing MDR-TB cases, notably following the results of the so-called Bangladesh study, which demonstrated that a CZM-containing regimen can cure such resistant cases in 9 to 12 months (5). Grosset et al. demonstrated the substantial benefit of adding CZM to second-line regimens in mice infected with isoniazid-resistant strains of M. tuberculosis (6). Furthermore, CZM was recently demonstrated to reduce the duration of TB in a mouse model of TB (7). However, CZM use is hampered by its common side effects, in particular skin discoloration, caused by its long half-life and extremely high lipophilicity (8). Through medicinal chemistry, new CZM analogues were synthesized, and these demonstrated equivalent or better efficacy than CZM in a murine model of TB with reduced lipophilicity thus reducing expected side effects (9). Although the exact mechanism of action of CZM is not entirely understood, it was elegantly demonstrated in Mycobacterium smegmatis that CZM is a prodrug, which is reduced by type 2 NADH-quinone oxidoreductase (NDH-2) and releases reactive oxygen species upon spontaneous reoxidation by O₂ (10). CZM is believed to compete with menaquinone (MK-4; vitamin K₂), the sole quinone present in mycobacteria and a key electron acceptor, for its reduction by NDH-2. MK-4, also known as menatetrenone (C₃₁H₄₀O₂; molecular weight, 444.65), consists of a quinone ring linked to a chain of four isoprenoid groups.

Benzothiazinones (BTZs) are an extremely potent class of novel antimycobacterials that act by blocking the synthesis of decaprenyl-phospho-arabinose, the precursor of the arabinans in the mycobacterial cell wall (11). The lead compound BTZ043 was demonstrated to be fully compatible with all the other approved or experimental TB drugs tested (12). Interestingly, BTZ043 and 2-piperazino-benzothiazinone 169 (PBTZ169), the preclinical drug candidate, were shown to act synergistically in vitro with bedaquiline (BDQ), an ATP synthase inhibitor (13). Compared to BTZ043, PBTZ169 has improved potency, safety, and efficacy in zebrafish and mouse models of TB, and highly encouraging results were obtained against chronic murine TB when PBTZ169 was administered in combination with BDQ and/or pyrazinamide (13).

In the present study, we first evaluated the activity of CZM against actively growing and nonreplicating bacteria in vitro as well as the effect of MK-4 supplementation on CZM activity. The mode of action of CZM is related to that of BDQ, since the electron transfer chain is coupled to ATP synthesis to produce energy at the plasma membrane level. We, therefore, tested the interaction profile between CZM and BTZ in vitro before evaluating the efficacy of a PBTZ169-CZM combination in a murine model of chronic TB.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

M. tuberculosis strains H37Rv and 18b were grown at 37°C with shaking in 7H9 broth (Difco) supplemented with Middlebrook albumin-dextrose-catalase enrichment, 0.2% glycerol, 0.05% Tween 80, and, in the case of 18b, 50 μg/ml streptomycin (STR) or on solid Middlebrook 7H10 medium (Difco) supplemented with 0.5% glycerol, Middlebrook oleic acid-albumin-dextrose-catalase (OADC), and, in the case of 18b, 50 μg/ml STR. Nonreplicating streptomycin-starved 18b cultures (SS18b) were generated as previously described (14).

Drugs and chemicals.

CZM, isoniazid (INH), and menaquinone (MK-4) were purchased from Sigma-Aldrich. Experimental drugs were provided by K. Andries (BDQ) and V. Makarov (BTZ043, PBTZ169). All the drugs and MK-4 were dissolved in dimethyl sulfoxide, except INH, which was dissolved in water. Drugs for the in vivo experiment were prepared as follows: PBTZ was prepared in 0.5% carboxymethyl cellulose (CMC) at pH 3.0, as acidification increases PBTZ solubility (13), after grinding in a mortar and CZM was prepared in acidified CMC with 0.4% Tween 80. We used a standardized pH for drug preparations within the same experiment to reduce the risk of variation. Drug solutions were prepared weekly and stored at 4°C.

Drug activity measurement by REMA.

Compound activity and effect of MK-4 supplementation were evaluated by the resazurin reduction microplate assay (REMA) as previously described in 7H9 medium (12). When applicable, growth medium was supplemented with MK-4 at the concentrations indicated. Briefly, bacterial stocks of M. tuberculosis H37Rv were generated from midlog cultures and frozen at −80°C to standardize the inoculum. Frozen aliquots of tubercle bacilli were thawed and diluted to an optical density at 600 nm (OD₆₀₀) of 0.0025 and added to the plates containing drug dilutions to obtain a total volume of 300 μl (48-well plate) or 100 μl (96-well plate). Plates were incubated for 6 days at 37°C before addition of resazurin (0.025%, wt/vol to 1/10 of well volume). After overnight incubation, fluorescence of the resazurin metabolite resorufin was determined (excitation at 560 nm and emission at 590 nm, measured by using a Tecan Infinite M200 microplate reader). The MIC was defined as the lowest concentration preventing resazurin turnover from blue to pink and confirmed by the level of fluorescence measured by the microplate reader. M. tuberculosis SS18b cultures were prepared as previously described and diluted to an OD₆₀₀ of 0.1. The effects of MK-4 supplementation on the REMA dose-response curves were investigated using medium that was supplemented with 0, 10, 100, or 1,000 μM MK-4.

Evaluation of compounds alone or in combination against replicating M. tuberculosis H37Rv or nonreplicating SS18b using CFU assays.

As described above, bacteria were incubated in 7H9 liquid medium for 7 days in the presence of combinations of compounds at their respective MICs, or fractions thereof for H37Rv (concentrations used in individual REMA assays), and inhibitory or subinhibitory concentrations for SS18b. Dilutions were plated on a solid medium (supplemented 7H10, with STR for 18b), and CFU counts were determined after 3 to 4 weeks of incubation at 37°C.

In vivo evaluation of the PBTZ-CZM combination in a murine model of chronic TB.

Female BALB/c mice, aged 5 to 6 weeks, were obtained from Charles River Laboratories (Lyon, France). The in vivo antimicrobial activities of the two drugs alone and in combination were assessed in the chronic model of TB by gavage 6 days a week for 4 weeks. Infection was established using a low-dose aerosol (∼200 CFU) generated by a custom-built aerosol exposure chamber (Mechanical Engineering Shops, University of Wisconsin, Madison). Experiments were approved by the Swiss Cantonal Veterinary Authority (authorization no. 2658). Drug treatment began 4 weeks after infection at the dose of 25 mg/kg of body weight for PBTZ169 and 20 mg/kg for CZM. Control and treated mice were sacrificed, the lungs and spleens homogenized, and dilutions plated on 7H10 agar enriched with 10% OADC and supplemented with cycloheximide (10 μg/ml), ampicillin (85 μg/ml), and 0.4% weight per volume activated charcoal (Sigma) to prevent compound carryover (15).

Statistical analysis.

CFU counts were log₁₀ transformed before analysis as mean log₁₀ CFU ± standard deviation (SD) and compared using Student's t test in Prism version 5.0 (GraphPad).

RESULTS

Effect of menaquinone supplementation on CZM activity against replicating and nonreplicating bacteria.

We measured the activities of CZM, BDQ, and INH against M. tuberculosis H37Rv by REMA fluorescence (Fig. 1) in the presence or absence of MK-4. CZM, BDQ, and INH had MICs in normal 7H9 medium of ∼0.5, 0.1, and 0.2 μg/ml, respectively. The effect of MK-4 varied considerably in the three assays. For INH, a negative control, we detected no change in activity at the three MK-4 concentrations tested but did note a slight increase in absolute levels of fluorescence. CZM and BDQ activities were significantly reduced upon MK-4 addition. The MIC for each drug increased by around 10-fold in this assay in the medium containing the highest concentration of MK-4 (1,000 μM). BDQ activity was not affected by 10 μM MK-4, but its MIC increased at the higher concentrations used. We noticed an inverse correlation between CZM activity and MK-4 levels, at the three concentrations tested (10, 100, and 1,000 μM). Assuming that the levels of cofactors like MK-4 are critical to maintain oxidative metabolism in the nonreplicating state, we repeated the previous assay with CZM using the SS18b model. Here again, CZM activity on nonreplicating bacteria was impacted by MK-4 supplementation, as noted in the REMA assay (Fig. 2A), where the effect of 1 μg/ml CZM in plain medium was suppressed by a high concentration of MK-4. To confirm this result, we analyzed the effect of MK-4 supplementation using the number of SS18b CFU as a growth readout (Fig. 2B). The activity of CZM against nonreplicating cells was partially neutralized by MK-4. CZM at 1 μg/ml had essentially no activity against SS18b in an MK-4-rich medium, and even at 10 μg/ml its activity was greatly reduced when MK-4 was present. The same trend was observed with actively growing H37Rv cells, as assessed by the CFU assay (Fig. 2C), where a difference of 3.6 log units in CFU counts was seen after exposure to CZM at 5 μg/ml in the absence or presence of 1,000 μM MK-4. There were no significant differences between the untreated controls, with or without MK-4.

FIG 1.

Effect of menaquinone supplementation on clofazimine (CZM), BDQ, and isoniazid (INH) against M. tuberculosis H37Rv by REMA. Cells were grown in a normal 7H9 medium (no MK-4) or in medium supplemented with increasing concentrations of menaquinone (10, 100, and 1,000 μM MK-4). REMA results are presented as mean ± SD values of triplicates. Drug concentrations are in micrograms per milliliter.

FIG 2.

Effect of menaquinone supplementation on clofazimine (CZM) activity against nonreplicating M. tuberculosis 18b cells (A, B) and actively growing M. tuberculosis H37Rv (C). (A) Cell activity assessed by REMA with increasing concentrations of menaquinone. Results are average values of duplicates. (B) 18b cells were plated after 7 days of drug exposure with (black) or without (light gray) MK-4 (1,000 μM) for CFU count assessment. (C) Cells plated the same as in (B), but with actively growing M. tuberculosis H37Rv cells. CZM concentrations are in micrograms per milliliter. ***, P < 0.0005; #, CFU count below the limit of detection in this assay (<200 CFU); ns, no statistical significance; NT, nontreated control.

Synergistic interaction between CZM and BTZ.

Since CZM and BDQ each interfere with cellular respiration and their activities are affected by MK-4, we hypothesized that there might also be a synergistic interaction between CZM and BTZ as was previously found between BTZ and BDQ. To test this possibility, we evaluated the interaction between BTZs and CZM against M. tuberculosis H37Rv by CFU determination (Fig. 3A and B). We used compounds at their individual MICs in REMA: 1.5, 0.4, and 125 to 150 ng/ml for BTZ043, PBTZ169, and CZM, respectively, or fractions thereof. The data in Fig. 3A show that the combination of 0.4 ng/ml of BTZ043 (one-quarter of its MIC) and 37.5 ng/ml CZM (around one-quarter of its MIC), each of which had no impact on bacterial growth alone, show clear bacteriostatic activity in combination (mean CFU count was equivalent to day 0). The same behavior was noticed between PBTZ169 and CZM, demonstrating that this was a compound class effect. The activity of 31.3 ng/ml of CZM (around one-quarter of its MIC) was significantly enhanced when combined with 0.1 ng/ml PBTZ169 (one-quarter of its MIC), which had no impact on its own, thus confirming the in vitro synergy.

FIG 3.

Synergistic interaction in vitro between clofazimine (CZM) and two benzothiazinones, BTZ043 (A) and PBTZ169 (B). We tested BTZ043, PBTZ169, and CZM effects alone or in combination, (A) BTZ043 and CZM and (B) PBTZ169 and CZM, on the viability of M. tuberculosis H37Rv. Following 7 days of incubation, bacteria were plated to determine CFU counts. The nontreated bacteria (NT) were also plated on days 0 and 7. **, P < 0.005.

Synergistic interaction between BTZ and BDQ or CZM against nonreplicating bacteria.

PBTZ169, like other cell wall inhibitors, is poorly active against nonreplicating bacteria (14). We evaluated by the CFU assay the effect of the two synergistic combinations, PBTZ-BDQ (13) and PBTZ-CZM (Fig. 3), against actively growing M. tuberculosis H37Rv and SS18b in vitro (Fig. 4). The two PBTZ169 concentrations used, 250 and 62.5 ng/ml, each of which had no impact on bacterial activity on their own, significantly improved the effect of BDQ or CZM, each at 250 ng/ml. These synergistic interactions against dormant bacilli were reproducible, and results were consistent between duplicates and when BTZ043 was used instead of PBTZ169 (data not shown).

FIG 4.

In vitro interactions between PBTZ169 and BDQ (A) and between PBTZ169 and clofazimine (CZM) (B) against nonreplicating M. tuberculosis 18b. Streptomycin-starved 18b cells were exposed to selected dilutions of the drugs for 7 days. The nontreated bacteria (NT) were also plated on day 0 and on day 7. Drug concentrations are in nanograms per milliliter. *, P < 0.05.

Effect of menaquinone supplementation on the synergy between CZM and PBTZ169.

We assessed by CFU counts the effect of MK-4 supplementation on the synergy between PBTZ169 and CZM (Fig. 5). In an MK-4-rich medium (1,000 μM), the cidal effects of PBTZ169 (0.8 ng/ml ≈ MIC in an MK-4-rich medium) (see Fig. S1 in the supplemental material) or CZM (1,000 ng/ml) alone were not significantly improved when the drugs were combined. The synergy between CZM and PBTZ169 is, therefore, lost when the growth medium is saturated with MK-4.

FIG 5.

In vitro interactions between PBTZ169 and clofazimine (CZM) against M. tuberculosis H37Rv in a menaquinone-rich medium. The effects of selected concentrations of PBTZ169 and CZM alone and in combination on the viability of M. tuberculosis H37Rv were assessed after 7 days exposure in a medium containing 1,000 μM menaquinone. Drug concentrations are in nanograms per milliliter. The nontreated bacteria (NT) were also plated on days 0 and 7. ns, nonspecific.

Combination study of PBTZ169 with CZM in a mouse model of chronic TB.

We assessed the synergistic combination discovered in vitro in the murine model of chronic TB after low-dose aerosol infection (Fig. 6). PBTZ169 (25 mg/kg) and CZM (20 mg/kg) were tested alone, and the two drugs were tested together against M. tuberculosis H37Rv. PBTZ reduced the bacterial burdens in the lung and spleen by 1.8 ± 0.2 and 2.4 ± 0.2 log units, respectively, compared to the bacterial load before treatment (day 0). CZM activity was even more pronounced, with log unit reductions of 4.0 ± 0.3 for the lung and 2.7 ± 0.2 for the spleen CFU counts. The dual therapy of PBTZ and CZM in combination displayed promising killing since it reduced the numbers of bacilli in the lungs and in the spleen by 4.6 ± 0.2 and 4.2 ± 0.2 log units, respectively.

FIG 6.

In vivo efficacy study of PBTZ169 and clofazimine (CZM), alone and in combination, in a mouse model of chronic TB, compared with untreated controls (NT). PBTZ169 and CZM were administered 6/7 days at 25 and 20 mg/kg of body weight per day, respectively. 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 ± SD CFU from 5 mice per group.

DISCUSSION

Efficacious TB treatment exists, but it requires strict implementation strategies, universal access to drugs, and careful compliance over a prolonged treatment period, particularly for drug-resistant cases. We need innovative combination regimens comprising new molecules able to kill drug-susceptible and drug-resistant strains of M. tuberculosis while simultaneously decreasing treatment duration by targeting active and persistent tubercle bacilli (16).

The mechanism of action of CZM is not entirely clear. It was presumed to compete with menaquinone, the sole quinone cofactor in mycobacteria, for electrons carried by the flavin adenine dinucleotide (FAD) moiety of reduced NDH-2, although such competition has not yet been demonstrated to our knowledge (10). If the mechanism of action of CZM remains obscure, a new mechanism of resistance was recently identified. CZM and BDQ share cross-resistance due to overexpression of the MmpL5 efflux system, which presumably reduces the intracellular concentration of each drug (17, 18).

MK-4 biosynthesis is essential for mycobacterial growth, and selective inhibitors of the menaquinone biosynthetic enzyme, MenA (1,4-dihydroxy-2-naphthoate octaprenyltransferase), kill nonreplicating M. tuberculosis and act as indirect inhibitors of ATP synthesis (19). An F₄₂₀ (redox enzyme cofactor)-dependent mechanism against oxidative stress may also be partly suppressed by menadione addition in M. tuberculosis (20). Isoniazid, moxifloxacin, and CZM were shown to elevate oxidative stress, and F₄₂₀-deficient mutants were hypersensitive to these molecules. In our work, we show that INH activity, as measured by REMA, is not affected by MK-4 supplementation, unlike CZM (Fig. 1). These microbiological findings suggest that the proton gradient and subsequent ATP synthesis may be influenced by MK-4 levels, thus validating the notion that MK-4 can suppress CZM activity in a dose-dependent manner and confirming the hypothesis of Yano et al. (10). Bacterial inhibition by CZM may be rescued by increasing MK-4 levels. CZM, and to some extent BDQ, thus behave similarly to Ro48-8071, a lipophilic amine inhibitor of MenA, whose activity was partly suppressed by 400 μM MK-4 (21). However, the antagonistic effect of MK-4 against CZM activity may not be of clinical relevance, as the concentrations used in the present work exceed the physiological levels in humans (22).

For an optimal regimen, benzothiazinones should be combined with chemotherapeutic agents able to target latent bacilli, such as BDQ, that inhibit the ATP synthase required to maintain mycobacterial viability during dormancy (23). In addition, the residual metabolism of dormant M. tuberculosis presumably requires some de novo RNA and protein synthesis for survival during this phase or for reactivation (24). Given the sensitivity of nonreplicating bacteria to inhibitors of these processes (14, 25), combining PBTZ169 and other drugs active against nonreplicating bacilli, such as BDQ, CZM, oxazolidinones, or pyrazinamide, offers an attractive foundation for a new TB regimen. In the present study, the combination of PBTZ169 and CZM was extremely potent in a murine model of chronic TB, and similar results are to be expected with the newer riminophenazines. These highly encouraging results are likely to be reproduced when combining PBTZ169 with other compounds affecting the electron transfer chain, such as the newly identified imidazopyridine amide Q203 (26). The possibility of combining menaquinone synthesis inhibitors with DprE1 inhibitors for putative synergy has already been raised (27).

The synergy between BTZ and BDQ was initially explained through weakening of the cell wall by DprE1 inhibition leading to better penetration of the ATP synthase inhibitor and improved access to its target (12, 13). Indeed, in support of this explanation, in vitro synergy was also reported between BDQ and the ethylenediamine SQ109 (28), whose primary target is the transmembrane protein MmpL3 required for trehalose monomycolate transport, the loss of which weakens the cell wall (29). This explanation for synergy appeared less likely when another mechanism of action of SQ109 was revealed, namely, inhibition of menaquinone biosynthesis (30).

Regarding the newly identified synergy between BTZ and CZM, this may also arise through some cell wall damage caused by subinhibitory concentrations of DprE1 inhibitors and improved penetration of CZM even in the SS18b latency model where a synergistic interaction between PBTZ and BDQ or CZM was detected. This may indicate that, although nonreplicating, dormant cells require some cell wall maintenance and retain limited sensitivity to cell wall inhibitors. For instance, INH monotherapy is known to prevent reactivation in latent TB infection, where M. tuberculosis is thought to persist in a dormant state (31, 32).

Another explanation for the in vitro synergy between BTZ and compounds affecting membrane potential stems from the enzymatic activity of DprE1 itself (33). DprE1 catalyzes the first step in the conversion of decaprenylphosphoryl-ribofuranose to decaprenylphosphoryl-arabinose, the sole precursor of the arabinan of the mycobacterial cell wall (34). This FAD-dependent process requires an electron acceptor for enzyme turnover, and reoxidation of reduced flavin adenine dinucleotide (FADH₂) by several electron acceptors was demonstrated in vitro, in particular by MK-4 (33). Therefore, loss of DprE1 activity, due to binding of BTZ, might result in fewer reducing equivalents from FADH₂ entering the electron transfer chain and thus enhance the effects of CZM and BDQ. Alternatively, by disrupting the proton gradient, CZM or BDQ may indirectly prevent reoxidation of FADH₂ in DprE1, hence improving its enzymatic inhibition by BTZ. It should be noted that the MIC of BTZ increased 4-fold in MK-4-rich medium (see Fig. S1 in the supplemental material) where the synergy between CZM and PBTZ169 was also lost (Fig. 5). Finally, whatever the explanation, the promising activity of the PBTZ169 and CZM combination is highly encouraging for the design of innovative TB regimens for humans, especially against drug-resistant strains.