Combinations of R207910 with Drugs Used To Treat Multidrug-Resistant Tuberculosis Have the Potential To Shorten Treatment Duration

Abstract

The objective of the present study was to identify the optimal R207910-containing regimen to administer to patients who cannot receive rifampin (RIF) and isoniazid (INH) because of multidrug-resistant tuberculosis (MDR-TB), concomitant use of antiretroviral drugs, or toxicity. Mice were infected intravenously with 5 × 10⁶ CFU of the H37Rv strain and treated five times per week with R207910 alone or various combinations of R207910 with the second-line drugs amikacin (AMK), pyrazinamide (PZA), moxifloxacin (MXF), and ethionamide (ETH). All R207910-containing regimens were significantly more active than the non-R207910-containing regimens after 1 month of therapy. When given for 2 months, R207910 alone was more active than the WHO standard first-line regimen RIF-INH-PZA. When R207910 was combined with second-line drugs, the combinations were more active than the currently recommended regimen of MDR-TB AMK-ETH-MXF-PZA, and culture negativity of both the lungs and spleen was reached after 2 months of treatment in almost every case.

R207910 (TMC207 or compound J), a diarylquinoline, is a new compound exhibiting a completely new mode of action (inhibition of ATP synthase) against mycobacteria (1). The compound is being developed in phase IIa trials for the treatment of active tuberculosis (TB) as TMC207. In mice, it has remarkable in vivo activity against Mycobacterium tuberculosis (1), M. leprae (10), and M. ulcerans (11). In the established murine model of TB, compound J on its own is as active as the standard WHO regimen of RIF-INH-PZA, the triple association of rifampin, isoniazid, and pyrazinamide. Furthermore, when compound J was added to RIF-INH-PZA, INH-PZA, or RIF-PZA, the lungs of mice harboring 5.94 log₁₀ CFU at the start of treatment became culture negative after just 2 months of treatment (1).

This powerful anti-TB activity of compound J raises the possibility of developing an anti-TB combination therapy without INH and RIF. INH and RIF are limited in TB treatment because of development of resistance, toxicity, hypersensitivity, and/or drug-drug interactions. Until now, when clinicians have been unable to use INH and RIF, so-called second-line drugs such as aminoglycosides (especially AMK), fluoroquinolones (moxifloxacin), ethionamide, para-aminosalicylic acid, cycloserine, and more recently linezolid have been used in various combinations.

In situations where INH and RIF cannot be used, the World Health Organization (WHO) currently recommends the use of a regimen including AMK, ethionamide, a fluoroquinolone (such as moxifloxacin), and PZA. This combination (AMK-ETH-MXF-PZA) is, however, less active than the standard RIF-INH-PZA regimen in the established mouse model of TB. Despite its better activity over RIF-INH-PZA after 2 months of treatment, 9 months of therapy are needed to reach culture negativity of both the lungs and spleen, while only 6 months are needed with the RIF-INH-PZA regimen (20).

In the present study, we examined the impact of combining compound J with drugs in the WHO-recommended AMK-ETH-MXF-PZA regimen in the established mouse model of TB and we compared the efficacy of the resulting combinations to that of the standard WHO RIF-INH-PZA regimen.

MATERIALS AND METHODS

Antimicrobial agents.

The compounds used were purchased from various manufacturers. INH was from Laphal (Allauch, France), RIF and PZA were from Aventis (Antony, France), AMK was from Bristol-Myers Squibb (La Défense, France), MXF was from Bayer (Puteaux, France), and ETH was from Sigma (France). Compound J was synthesized by Johnson & Johnson (Beerse, Belgium).

M. tuberculosis strain.

The drug-sensitive H37Rv strain of M. tuberculosis was grown on Löwenstein-Jensen medium. Colonies were subcultured in Dubos broth (Diagnostics Pasteur, Paris, France) for 7 days at 37°C. The turbidity of the resulting suspension was adjusted with normal saline to match that of a standard 1-mg/ml suspension of M. bovis BCG and was further diluted with normal saline to obtain a 0.2-mg/ml suspension for mouse inoculation. The MICs for the H37Rv strain were 0.25 μg/ml RIF, 0.06 μg/ml INH, 1.0 μg/ml AMK, 0.5 μg/ml MXF, 0.5 μg/ml ETH, 0.03 μg/ml compound J (all of which were determined on 7H11 agar medium), and 10 μg/ml PZA (determined on Löwenstein-Jensen medium) at pH 5.5.

Infection of mice.

Two hundred thirty female 4-week-old outbred Swiss mice, purchased from the Janvier Breeding Center (Le Genest Saint-Isle, France), were inoculated in the tail vein with 0.5 ml of a bacterial suspension containing approximately 5 × 10⁶ CFU of M. tuberculosis H37Rv.

Chemotherapy.

Following infection, mice were randomly allocated to six control groups and five test groups (Table 1), each of 20 to 30 mice. The first group was a negative control group, in which mice were infected but left untreated; mice in the second and third groups were treated with compound J and MXF (compound M) monotherapy for 2 months. The fourth and fifth groups were positive control groups; in the fourth group, mice were treated with the standard regimen for drug-susceptible TB, i.e., 2 months of the combination RIF-INH-PZA (13, 22), and mice in the fifth group received the standard WHO RIF-INH-PZA regimen supplemented with compound J (1). Mice in the sixth group were treated with the multidrug-resistant TB (MDR-TB) regimen recommended by the WHO (3), in which ofloxacin was replaced with MXF. The resulting AMK-ETH-MXF-PZA regimen is currently the best available option for the treatment of MDR-TB (20). Mice in the remaining groups were treated for 2 months with AMK and PZA supplemented, respectively, with ETH; MXF; ETH and compound J; MXF and compound J; or ETH, MXF, and compound J.

TABLE 1.

Experimental design used in this study

Group	No. of mice
	Day 13	Day 0	1 mo	2 mo
Untreated	10	20a
Compound J aloneb			10	10
M alone			10	10
RHZ			10	10
RHZJ			10	10
AEtMZ			10	10
AEtZ			10	10
AMZ			10	10
AEtZJ			10	10
AMZJ			10	10
AEtMZJ			10	10

^aTen mice for sacrifice at days 0 and 10 were left for mortality.

^bDosages used: compound J, 25 mg/kg; moxifloxacin (M), 100 mg/kg; rifampin (R), 10 mg/kg; isoniazid (H), 25 mg/kg; pyrazinamide (Z), 150 mg/kg; amikacin (A), 150 mg/kg; ethionamide (Et), 50 mg/kg.

Treatment was initiated 2 weeks after infection in order to achieve a larger bacterial population and was administered 5 days weekly. AMK was diluted in normal saline and given by subcutaneous injection. Compound J was prepared monthly in a hydroxypropyl-β-cyclodextrin solution, as such a solution was shown to be chemically stable for a month when kept at 4°C, and given orally by gavage. The remaining drugs were also given orally. Suspensions in distilled water containing 0.05% agar were prepared weekly and stored at 4°C.

The drugs were administered in the following dosages (in milligrams per kilogram per day): compound J, 25; RIF, 10; INH, 25; PZA, 150; MXF, 100; AMK, 150; ETH, 50. Based on the areas under the concentration-time curves, these dosages, which are similar to those used in previous experiments (8, 9, 13, 14, 15, 19), were chosen as equipotent with the usual dosages administered to humans. AMK was given at 150 mg/kg based on the body area distribution, which is 10 times higher in mice than in humans and also based on previous efficacy studies with mice.

Assessment of infection and treatment.

To provide baseline values, 10 infected and 10 untreated mice were sacrificed on days 1 and 14 after infection (days −13 and 0 in relation to the initiation of treatment). In the drug-treated groups, sacrifices were carried out after 1 and 2 months of treatment. The severity of infection and the effectiveness of treatments were assessed by survival rate, spleen weight, gross lung lesions (scored from 0 to ++, the latter referring to a lung that was extensively occupied by tubercles), and numbers of CFU in the lungs and spleen. At days −13 and 0 and after 1 month of treatment, the number of CFU in the spleen and lungs were determined by plating three serial 10-fold dilutions of homogenized suspensions onto triplicate Löwenstein-Jensen slants. After 2 months of treatment, the entire suspension prepared from each individual organ of the compound J-containing groups, expected to contain only a few bacilli, was plated without dilution on 15 Löwenstein-Jensen slants. Results of the cultures were recorded after incubation at 37°C for 6 weeks. The bactericidal effect of the treatment was defined as a significant decrease in the mean number of CFU in the treated group compared to the pretreatment value.

Statistical analysis.

Results were analyzed by the Student t test and Fisher's exact probability calculation. Differences were considered significant at the 95% level of confidence.

RESULTS

Survival rate.

Untreated mice all died between days 21 and 28 after infection, while all treated mice survived the infection but a few died from gavage accidents.

Mean spleen weight.

Before treatment, the mean spleen weight (± the standard deviation) increased significantly from 112 ± 14 mg the day after infection to 631 ± 121 mg 2 weeks postinfection (P < 0.05), when all of the treatments were started, indicating active replication of bacilli between days 1 and 14. After 1 month of therapy, all of the treated groups displayed significantly decreased mean spleen weights in comparison to the spleen weights at the start of treatment (P < 0.05). Compound J-treated groups displayed mean spleen weights similar to those of groups not treated with compound J. An additional month of treatment did not further reduce the mean spleen weights of the different treatment groups.

Gross lung lesions.

At the time treatments were started, all mice harbored hundreds (++) of gross lung lesions (results not shown). One month of treatment with MXF alone was not able to reduce the number of gross lung lesions, but treatment with compound J alone did (from ++ to +). All combinations were able to reduce the number of gross lung lesions from hundreds (++) to fewer than 10 lesions (tubercles) (+) during the first month of therapy.

An additional month of treatment with compound J alone was not able to cure the gross lungs lesions; they remained at the level seen at the end of 1 month of treatment. MXF alone given for 2 months was able to decrease the lung lesion score from ++ to +. All of the combinations of drugs were able to decrease the number of gross lung lesions in comparison with 1 month of treatment but only if compound J was included in the combination (from + to 0).

Enumeration of CFU in the spleens.

At the beginning of treatments, there were 6.50 ± 0.23 log₁₀ CFU in the spleens of untreated mice (Table 2). After 1 month of treatment, monotherapy with compound J or MXF was able to reduce significantly the number of bacilli in the spleen in relation to pretreatment values (P < 0.05). MXF alone was able to kill 1.5 log₁₀ CFU, while compound J alone killed almost 4 log₁₀ CFU of M. tuberculosis. Regimens that did not include compound J were able to kill 2 to 3 log₁₀ CFU of M. tuberculosis. All of the compound J-containing regimens were significantly more active than non-compound J-containing regimens (P < 0.05), killing 4 to 5 log₁₀ CFU. After 2 months of treatment, compound J alone and MXF alone had killed >5 and 3 log₁₀ CFU, respectively. Combinations excluding compound J killed 3.7 to 4.9 log₁₀ CFU. Compound J-containing regimens were significantly more active than those that did not contain compound J (P < 0.05). The combinations AMK-ETH-PZA-J and RIF-INH-PZA-J both killed 6.4 log₁₀ CFU, and the combinations AMK-MXF-PZA-J and AMK-ETH-MXF-PZA-J killed all of the bacilli present at the beginning of treatment.

TABLE 2.

CFU counts (log₁₀) in the spleens and lungs of mice after 1 and 2 months of treatment and proportions of mice displaying negative cultures

Regimen	Mean no. of CFU ± SD (no. of mice with negative cultures/total)
	Spleen at 1 mo	Spleen at 2 mo	Lungs at 1 mo	Lungs at 2 mo
Untreated, day 0	6.5 ± 0.2		5.9 ± 0.5
Compound J alone	2.6 ± 1.3	1.2 ± 0.5 (0/8)	2.9 ± 0.9	0.2 ± 0.3 (6/8)
Ma alone	5.0 ± 0.3	3.5 ± 0.4 (0/10)	4.2 ± 0.3	2.9 ± 0.6 (0/10)
RHZ	4.5 ± 0.3	1.9 ± 0.5 (1/10)	3.7 ± 0.4	1.0 ± 0.5 (0/10)
RHZJ	1.9 ± 0.31	0.1 ± 0.2 (4/10)	1.8 ± 0.4	0 ± 0 (10/10)
AEtMZ	3.2 ± 0.5	1.6 ± 0.4 (1/10)	2.9 ± 0.2	0.1 ± 0.1 (5/10)
AEtZ	4.0 ± 0.3	2.8 ± 0.3 (0/10)	3.7 ± 0.2	1.2 ± 0.3 (0/10)
AMZ	3.6 ± 0.2	1.9 ± 0.5 (0/10)	3.4 ± 0.3	0.8 ± 0.6 (0/10)
AEtZJ	1.2 ± 0.2	0.1 ± 0.1 (7/9)	0.2 ± 0.3	0 ± 0 (9/9)
AMZJ	1.2 ± 0.2	0 ± 0 (8/8)	0.2 ± 0.3	0 ± 0 (8/8)
AEtMZJ	1.2 ± 0.3	0 ± 0 (8/8)	0.5 ± 0.4	0 ± 0 (8/8)

^aFor an explanation of the abbreviations used, see Table 1, footnote b.

In terms of the proportion of negative cultures, all of the compound J-containing regimens, except the RIF-INH-PZA-J regimen, were significantly more potent in rendering spleens culture negative than the regimens that did not include compound J (P < 0.05). AMK-ETH-MXF-PZA-J was significantly more active than AMK-ETH-MXF-PZA in reducing CFU counts (P = 0.0001) and in rendering spleens culture negative (P = 0.0001).

Enumeration of CFU in the lungs.

At the beginning of the treatments, there were 5.94 ± 0.51 log₁₀ CFU in the lungs of untreated mice (Table 2).

After 1 month of treatment, the compound J and MXF monotherapy regimens were able to reduce the number of bacilli in the lungs by 3.0 and 1.7 log₁₀ CFU of M. tuberculosis, respectively (P < 0.05). Combinations not containing compound J were able to kill between 2.2 and 3.0 log₁₀ CFU of M. tuberculosis, and compound J-containing combinations killed between 4.1 and 5.7 log₁₀ CFU. All compound J-containing regimens were significantly more active than the non-compound J-containing regimens (P < 0.05).

After 2 months of treatment, compound J alone and MXF alone had killed 5.7 and 3.0 log₁₀ CFU of M. tuberculosis, respectively. Combinations excluding compound J killed between 4.9 and 5.8 log₁₀ CFU, while combinations including compound J were able to kill all of the bacilli present at the beginning of treatment (Table 2). Compound J-containing regimens were significantly more active than those that did not contain compound J (P < 0.05), except for the AMK-ETH-MXF-PZA regimen (P > 0.05).

In terms of the proportion of negative cultures, compound J-containing regimens were significantly more potent in rendering lungs culture negative than the regimens that did not include compound J (P < 0.05). In fact, AMK-ETH-MXF-PZA-J rendered 100% of the lungs culture negative while the AMK-ETH-MXF-PZA regimen rendered only 50% of the lungs culture negative (P = 0.036). However, the difference in CFU counts between these two regimens did not reach statistical significance because of the very small numbers of bacilli (0.1 versus 0.0 log₁₀).

DISCUSSION

The present study tried to identify the optimal compound J-containing combination to use when INH and RIF cannot be used in situations such as MDR-TB, drug toxicity, or interaction of RIF with antiretroviral drugs. The regimen containing AMK, PZA, MXF, and ETH is currently the most potent one against MDR-TB in the mouse model. This combination rendered lung and spleen cultures negative in a significantly shorter treatment duration (9 months) than the old anti-TB regimen consisting of streptomycin, INH, and ethambutol, which did not render the organs culture negative even after 12 months of treatment (unpublished data). However, use of the AMK-ETH-MXF-PZA regimen in humans is associated with two problems; i.e., it is less effective than RIF-INH-PZA, and hepatotoxicity related to the combination of PZA and ETH is frequently reported (21). Substituting a new potent drug for ETH could solve both problems.

Compound J as monotherapy proved to be as active as the standard RIF-INH-PZA regimen and as active as the WHO-recommended AMK-ETH-MXF-PZA regimen after both 1 and 2 months of treatment. Of course, compound J monotherapy would never be adopted as a treatment option because of the risk of selecting compound J-resistant mutants during therapy.

When comparing the efficacy of compound J to that of ETH in the AMK-MXF-PZA-J and AMK-ETH-MXF-PZA regimens, the former is more active both in terms of reducing CFU counts in the spleens and lungs and in terms of the proportions of mice displaying negative cultures after 2 months of therapy. This can be explained by the higher activity of compound J over ETH both in vitro and in vivo. The MICs of ETH against M. tuberculosis range between 0.5 and 4 mg/liter, and those of compound J range between 0.03 and 0.12 mg/liter. In infected mice, 50 mg/kg ETH killed 1.2 log₁₀ CFU in the lungs after 4 weeks of treatment (4) and 0.4 log₁₀ CFU after 4 weeks in another study (12). In a study where mice were infected with an MDR-TB strain and treated for 4 weeks, 100 mg/kg ETH killed 0.9 and 0.4 log₁₀ CFU in the spleen and in the lungs, respectively (5). Compound J, on the other hand, killed 3.0 log₁₀ CFU in the lungs after 1 month of treatment and more than 5.0 log₁₀ CFU in the lungs after 2 months of treatment with a 25-mg/kg dose (1). Compound J therefore seems much more potent than ETH, both in vitro and in the mouse model.

When comparing MXF to ETH in the AMK-MXF-PZA and AMK-ETH-PZA regimens, the contribution of MXF to bactericidal activity in terms of lung and spleen CFU counts seems to be more important than the contribution of ETH (P < 0.05). However, the ability to render organs culture negative is very poor for both regimens. When compound J is added to AMK-MXF-PZA and AMK-ETH-PZA, resulting in the AMK-MXF-PZA-J and AMK-ETH-PZA-J regimens, the difference in efficacy in terms of CFU counts disappears and almost all organs become culture negative after 2 months of treatment.

When comparing MXF to compound J in the AMK-ETH-MXF-PZA and AMK-ETH-PZA-J regimens, the latter is much more active both in terms of reducing CFU counts and in terms of reaching culture negativity after 2 months of treatment. This is not surprising, as the comparison between MXF and compound J as monotherapy indicates that compound J at a dose of 25 mg/kg is more bactericidal than MXF at a dose of 100 mg/kg (P < 0.05).

Combination regimens including compound J were more active than the standard RIF-INH-PZA or AMK-ETH-MXF-PZA regimen. When compound J was added to either AMK-ETH-MXF-PZA or AMK-MXF-PZA, the efficacy reached after 4 weeks was similar to that reached by AMK-ETH-MXF-PZA after 8 weeks. In addition, both lung and spleen cultures became culture negative after just 2 months of therapy. Historical data indicate that the AMK-ETH-MXF-PZA regimen needs 9 months to render both spleens and lungs culture negative (20). These observations suggest that compound J-containing MDR-TB regimens have the potential to shorten treatment duration by at least 50%, similar to what was seen when compound J was added to the first-line RIF-INH-PZA treatment regimen (1). The results of this experiment also suggest that addition of compound J may increase the potency of the currently recommended MDR regimen, which is known to be less efficient than the first-line regimen, to a level similar to that of the first-line regimen RIF-INH-PZA. The AMK-MXF-PZA-J regimen was as active as the AMK-ETH-MXF-PZA-J regimen and offers the potential to eliminate the need to include ethionamide. A regimen avoiding the use of aminoglycosides would even be more desirable because of their ototoxicity and nephrotoxicity and because of the risk of transmission of human immunodeficiency virus and hepatitis B and C among patients after injection of aminoglycosides.

Other studies omitting the use of INH, RIF, or both have been performed with the mouse model, but in none of them was culture negativity reached after 2 months of treatment. In a study using RIF-MXF-PZA, a regimen excluding INH, the lung CFU count after 2 months of treatment was 1.1 log₁₀, (a reduction of 5.7 log₁₀ CFU) (17). Use of Pa-824-MXF-PZA, a regimen excluding both INH and RIF and including the nitroimidazopyran Pa-824, reduced CFU counts in the lungs by 6.2 log₁₀ CFU but did not reach negativity since 1.5 log₁₀ CFU remained in the lungs at the end of treatment (18). In a study reported by Arora et al. (2), a combination of 25 mg/kg lupin compound, 100 mg/kg ethambutol, and 150 mg/kg PZA given 5 days per week for 2 months to mice infected with a susceptible strain of M. tuberculosis killed 3.5 log₁₀ CFU in the lungs and 3.9 log₁₀ CFU in the spleens, and the remaining CFU counts in the lungs and the spleens were 2.5 and 2.6 log₁₀ CFU, respectively. The same combination of drugs was tested in mice infected with an MDR strain of M. tuberculosis and killed 2.9 and 3.0 log₁₀ CFU in the lungs and in the spleens, respectively. The remaining CFU counts in the lungs and in spleens were 3.2 and 3.4 log₁₀, respectively. The most recent study by Matsumoto et al. (16) combined OPC-67683, a new nitro-imidazooxazole, with RIF and PZA, and the CFU counts remaining in the lungs were higher than 2 log₁₀ after 2 months of therapy.

A regimen containing AMK, ETH, MXF, PZA, and compound J appears to be the most promising for treatment of MDR-TB. Many studies showed that AMK was the best aminoglycoside for the treatment of TB in the mouse model (14) and ETH was the most active drug against MDR-TB in mice infected with an MDR strain compared to cycloserine, thiacetazone, para-aminosalicylic acid, capreomycin, and linezolid (5). MXF was the most active fluoroquinolone against M. tuberculosis in the mouse model (9) and displayed significant early bactericidal activity in TB patients (6). PZA is the most active drug against persistent bacilli (23), and compound J displayed important activity against both sensitive TB strains and MDR-TB strains in vitro (7) and has very powerful activity against sensitive M. tuberculosis in both acute and established murine TB models (1).

In conclusion, our study is the first one in which a regimen without isoniazid and rifampin was able to reach lung and spleen negativity within 2 months of therapy. The addition of compound J to combinations for treatment of drug-sensitive TB and MDR-TB appears to accelerate bacterial clearance and supports the hope for shortening treatment duration in humans, provided an equipotent dose proves to be safe for use in patients. As the present study did not allow estimation of the individual contributions of AMK and PZA, which were kept as a backbone in the MDR treatment regimens, further studies assessing various combinations of compound J with AMK, MXF, ETH, and PZA are needed to fully comprehend their respective contributions to their efficacy. In addition, relapse rates with the most promising combinations need to be determined to fully understand the potential of these new regimens.