|Home | About | Journals | Submit | Contact Us | Français|
The in vitro interactions of two new antitubercular drugs, SQ109 and TMC207, with each other and with rifampin (RIF) were evaluated. The combination of SQ109 with TMC207 (i) improved an already excellent TMC207 MIC for M. tuberculosis H37Rv by 4- to 8-fold, (ii) improved the rate of killing of bacteria over the rate of killing by each single drug, and (iii) enhanced the drug postantibiotic effect by 4 h. In no instance did we observe antagonistic activities with the combination of SQ109 and TMC207. Rifampin activates cytochrome P450 genes to reduce the area under the curve (AUC) for TMC207 in humans. The presence of RIF in three-drug combinations did not affect the synergistic activities of SQ109 and TMC207, and SQ109 also dramatically decreased the MIC of RIF. SQ109 was active by itself, and both its activity was improved by and it improved the in vitro activities of both RIF and TMC207.
Pulmonary tuberculosis (TB) is currently treated with a regimen of four drugs discovered before 1970: rifampin (RIF), isoniazid (INH), pyrazinamide (PZA), and ethambutol (EMB) or streptomycin (STR). These four drugs are administered for 2 months (the intensive phase), followed by the administration of RIF and INH for 4 to 7 months (the continuation phase), for a total of 6 to 9 months of treatment (1, 19). Because of the difficulty in eradicating Mycobacterium tuberculosis from tissues with the currently available drugs and the long duration of treatment for TB, many patients fail to take all their drugs or terminate their therapy early, and noncompliance with therapy has fueled the development of M. tuberculosis drug resistance (3). Multidrug-resistant TB (MDR-TB), defined as infection with M. tuberculosis isolates resistant to both RIF and INH, requires up to 24 months of treatment with a cocktail of less effective and more toxic second-line drugs to achieve a cure.
At least seven new antitubercular drugs are currently under evaluation in human clinical trials (18). The new drug candidates in development act on different metabolic or structural targets in M. tuberculosis, and all act on different targets than the current front-line TB drugs. SQ109 [N-geranyl-N′-(2-adamantyl) ethane-1,2-diamine] is a new diamine antitubercular drug candidate that interferes with cell wall synthesis in M. tuberculosis (9, 14). The exact mechanism of action (MOA) on the cell wall is not yet known. TMC207 (also known as R207910) is a diarylquinoline that specifically inhibits mycobacteria by inhibition of ATP synthesis (2). Individually, both SQ109 and TMC207 are bactericidal, effective against MDR-TB strains, and synergistic with (or can be synergistically enhanced by) other front-line anti-TB drugs (5, 8). SQ109 acts synergistically to improve the antitubercular action of INH and RIF (5), whereas TMC207 is synergistic with PZA (8). In humans, both drugs have a long half-life: SQ109 has a half-life of 60 h at 300 mg/day (unpublished data), and TMC207 has a half-life of 24 h at 400 mg/day (2). Therefore, both drugs have the potential for intermittent dosing. Of interest, the area under the curve (AUC) for TMC207 in humans was decreased by the activation of certain cytochrome P450 enzymes by RIF (10). Since SQ109 can lower by 4-fold the MIC of RIF, we were interested to see whether SQ109 in combination with TMC207 might mitigate the adverse drug-drug interaction that occurs with TMC207 in humans in the presence of RIF.
In this study we investigated the in vitro efficacy of combinations of TMC207 and SQ109 at killing M. tuberculosis, with and without the presence of RIF.
TMC207 was obtained from Tibotec Pharmaceuticals, Ltd. RIF and INH were purchased from Sigma-Aldrich. Current GMP SQ109 was synthesized and purified under contract for ongoing human clinical trials by the National Cancer Institute, NIH, Bethesda, MD.
The following mycobacterial strains maintained in our laboratory were used in this study: laboratory strain M. tuberculosis H37Rv, the M. tuberculosis H37Rv bacterial luciferase reporter strain (strain pSMT1) (17), and M. smegmatis MC2155. The mycobacteria were grown in 7H9 broth supplemented with 10% albumin-dextrose-catalase (ADC) and 0.05% Tween 80. Log-phase cultures incubated at 37°C on an orbital shaker were used in all assays.
Drug interactions were determined by checkerboard titration of TMC207 and SQ109, SQ109 and RIF, or TMC207 and RIF by a microdilution assay described previously (5, 7). In some experiments, we used three-drug combinations. In these three-drug studies, a standard checkerboard titration was performed with SQ109 and TMC207 and then different concentrations of RIF were added at one concentration per plate. The drugs were diluted in 100-μl volumes in 96-well microtiter plates (Nunc, Roskilde, Denmark). One drug was diluted vertically (rows A to G) and the second drug was diluted horizontally (columns 1 to 10) to obtain various combinations of the two drugs. The bottom row (row H) and column 11 in the plate contained individual drugs, and the last column (column 12) was the drug-free control. A log-phase culture of M. tuberculosis H37Rv was added (100 μl) to each well. The plates were placed in zip-lock bags and incubated at 37°C for 2 weeks before the results were read. The concentrations in the wells showing no visible growth were considered the inhibitory concentrations.
Fractional inhibitory concentrations (FICs) were calculated by use of the following formula to determine synergistic activity: FIC = MIC in combination/MIC alone. The fractional inhibitory index (ΣFIC) was determined and was calculated as the FIC of drug A + FIC of drug B. As defined in the publication detailing the assay, ΣFICs of ≤0.5 indicate synergistic activity, ΣFICs of ≥4.0 indicate antagonistic activity, and values in between indicate an additive interaction (7, 13). Although the term “additive” is used in the reference paper, the middle values (>0.5 and <4.0) could also be considered indifferent or could indicate that the two drugs act independently of each other. For the purposes of this paper, we assume that the terms “additive” and “indifferent” are interchangeable.
The numbers of viable M. tuberculosis isolates initially and at different times after incubation thereafter were determined by three methods that estimate killing by measuring metabolic inhibition or the actual reduction in the numbers of CFU: counting of the relative light units (RLUs) in a luminometer, determination of the growth index (GI) in a Bactec TB-460 reader (Becton Dickinson, Sparks, MD), and counting of the numbers of CFU on 7H11 agar plates. The rate of killing was determined by estimating the time that a drug or a drug combination eliminated >90% of the metabolic activity or >90% of the CFU in the initial inoculum.
In the RLU assay, we used M. tuberculosis pSMT1, which expresses bacterial (Vibrio harveyi) luciferase (17), an enzyme that catalyzes the oxidation of reduced flavin mononucleotide (FMNH2). The resulting long-chain fatty aldehyde emits a blue-green light at 490 nm. The reaction is as follows: FMNH2 + RCHO + O2 → FMN + RCOOH + H2O + light. After 1 week of growth, an M. tuberculosis pSMT1 suspension was centrifuged and the pellet was resuspended in 7H9 broth. The turbidity (λ600) was adjusted to 0.05. The suspension was distributed in 10-ml volumes in 50-ml tubes. SQ109, TMC207, and RIF, alone or in two- or three-drug combinations, were dispensed into the M. tuberculosis pSMT1 suspensions to give three increasing concentrations of the microdilution MIC of the individual drug, as indicated in Table Table4.4. A drug-free control growth tube was included in each experiment. At each time, the RLUs of quadruplicate 0.1-ml aliquots were determined (17).
Bactec 12B vials (Becton Dickinson, Sparks, MD) were injected with 0.1 ml of different concentrations of the drugs, followed by the addition of 0.1 ml of M. tuberculosis H37Rv suspensions adjusted to a λ600 of 0.1. A drug-free control group injected with 0.1 ml of diluent was included with each run. The Bactec vials were incubated at 37°C, and the GI was read daily in a Bactec TB-460 reader (Becton Dickinson, Sparks, MD). The drugs were tested at three increasing concentrations of their individual Bactec MICs, as detailed in Table Table55.
Different concentrations of drugs (microdilution MICs), alone and in combination, were prepared in 10-ml volumes of 7H9 broth in 50-ml tubes. A log-phase culture of M. tuberculosis H37Rv was adjusted to an optical density at 600 nm of 0.2, 0.1 ml was added to each tube, and the tubes were incubated on a shaker incubator. At different time points (days 0, 4, 8, and 16), 0.1- ml volumes of 10-fold dilutions of the M. tuberculosis suspensions were plated on replicate 7H11 agar plates and the plates were incubated at 37°C for up to 15 days. The numbers of colonies on replicate plates for each drug, drug combination, or concentration are reported as the mean log10 numbers of CFU at each time of assessment.
The postantibiotic effect (PAE) was determined as described previously (4). Briefly, 1.0 ml of an M. tuberculosis H37Rv suspension adjusted to a λ600 of 0.2 was added to 9 ml of prewarmed 7H9 broth. The drugs (TMC207 or SQ109 alone or SQ109 plus TMC207 or INH, each at a final concentration of 4 μg/ml) were added, and the tubes were incubated for 2 h at 37°C on a shaker. A drug-free growth control was also included. At the end of the incubation, suspensions were diluted 1/25 and 0.1 ml was injected into each of triplicate Bactec vials: the ensuing drug dilution was 1,000-fold in the Bactec vials. Drug carryover controls (which were used to ascertain the absence of any drug carryover effect) consisting of the same M. tuberculosis suspension used in the test and similar drug dilutions were mixed and immediately diluted 1/25, and 0.1 ml was injected into the Bactec vials. The Bactec vials were incubated at 37°C and read in a Bactec TB-460 reader every 24 h until the cumulative GI in the control vial reached 999. The PAE was calculated from the formula T − C, where T is the time required for the drug-exposed organisms to reach a cumulative GI of 100, and C is the time required for the controls to reach a cumulative GI of 100.
By the microdilution method, the MIC of SQ109 was 0.5 μg/ml, that of TMC207 was 0.25 μg/ml, and that of RIF was 0.0078 μg/ml. By use of the Bactec system, the MICs of these drugs against M. tuberculosis H37Rv were 0.25, 0.062, and 0.062 μg/ml, respectively. The MIC values were median values of triplicate experiments. The MIC of SQ109 or TMC207 was 1 or 2 dilutions lower by use of the Bactec system than by use of the microdilution method, but this difference was reversed with RIF. The MICs of the drugs were, however, comparable to the MICs published elsewhere (5, 9, 16) and were consistent with duplication within an assay type.
The activity of the combination of SQ109 with TMC207 against M. tuberculosis H37Rv was synergistic, with the ΣFICs being 0.375 in one experiment and 0.5 in the second experiment (Table (Table1).1). In each case, the MIC of SQ109 in the presence of TMC207 was improved 4-fold, whereas the MIC of TMC207 in the presence of SQ109 was improved 4-fold in one experiment and 8-fold in the second experiment. There was no change in either the TMC207 MIC or the RIF MIC when the two drugs were used in combination, and analysis suggested that the activity of the combination was additive or indifferent (ΣFIC, 2.0). On the other hand, the combination of SQ109 and RIF was synergistic (ΣFIC, 0.094): the presence of RIF caused the SQ109 MIC to be 16-fold lower than that obtained with SQ109 alone; the MIC of RIF in the presence of SQ109 was 30-fold lower than the RIF MIC when it was used alone. The synergistic activity of SQ109 and RIF has been remarkably consistent when the experiments performed by different investigators over different time periods (see reference 5 and Table Table22).
The activity of the combination SQ109 and TMC207 against M. smegmatis was also additive (ΣFIC, 1.0), and the MIC values changed only 2-fold for either drug (Table (Table1).1). The activity of the combination of SQ109 and RIF, however, against M. smegmatis was synergistic, with the ΣFIC being 0.132. The addition of RIF to SQ109 improved the SQ109 MIC 8-fold, and the addition of SQ109 to RIF improved the RIF MIC an astonishing 130-fold, rendering this drug effective against normally resistant mycobacteria at concentrations achievable in vivo.
No combination (SQ109-TMC207, SQ109-RIF, or TMC207-RIF) demonstrated antagonistic interactions in these studies.
To determine whether RIF improved the synergy between SQ109 and TMC207, we titrated TMC207 and SQ109 by checkerboard titration in microtiter plates, and different concentrations of RIF were added to individual plates: 0.5×, 0.2×, 0.1×, and 0.05× MIC RIF. Table Table33 shows the data for the three-drug combination from one of the two similar experiments. It should be noted that the experimental protocols used to assess the interactions of two drugs (7, 13) are not designed to simultaneously assess the interactions of three drugs. Had we seen marked additions to synergy, this would have been noteworthy. However, the interpretation of the results of these studies should be restricted to the detection of big changes from the data observed in the two-drug studies.
RIF at 0.05× and 0.1× MIC, concentrations that were not synergistic with SQ109, did not alter the synergy between SQ109 and TMC207 (Table (Table3,3, plates 3 to 5). In these plates, enhancement of the SQ109 MIC by TMC207 (2-fold) and enhancement of the TMC207 MIC by SQ109 (4-fold) were unaffected by the presence of RIF. At RIF concentrations of 0.5× and 0.2× MIC, the synergistic activity of SQ109 with TMC207 was not improved, but at TMC207 concentrations that were not synergistic with SQ109, synergy was evident with RIF (a 2-dilution, or 4-fold, decrease in the MIC of SQ109) (Table (Table3,3, plates 1 and 2). RIF did not contribute to the synergy between the combination of SQ109 and TMC207 that already existed (Table (Table3,3, plate 5). A two-drug control consisting of SQ109 and RIF showed good synergy, with ΣFIC being 0.094 (Table (Table22).
We conclude from the results of these studies that the activity of SQ109 is synergistically enhanced by TMC207 and RIF, that SQ109 synergizes the activity of TMC207 and RIF, and that TMC207 and RIF work fine together (additive or indifferent) when they are combined (Table (Table1).1). The addition of all three drugs together does not markedly improve or detract from the individual drug interactions (Table (Table33).
Since TMC207 affects mycobacterial ATP synthesis, we used a recombinant M. tuberculosis strain that expresses bacterial luciferase to assess the decreases in FMNH2 rather than an M. tuberculosis strain that expresses firefly luciferase to measure ATP. The rate of killing by each drug, alone or in combination, was determined by assessing the time required to kill ≥90% of the M. tuberculosis pSMT1 isolates in the original culture at time zero at a given concentration of drug. RLU values of ≤133 indicated ≥90% metabolic inhibition of the M. tuberculosis pSMT1 inoculum.
SQ109 by itself caused a ≥90% reduction in the numbers of RLUs by day 7 at all three concentrations tested (Table (Table4).4). The interaction of TMC207 with M. tuberculosis pSMT1 was dose dependent and more complicated: at 2× MIC, there was an initial rapid reduction in the numbers of RLUs that slowed and that did not reach >90% by day 12 but then increased to greater than the initial numbers of RLUs at day 16; at 4× and 8× MIC, however, the decline in the numbers of RLUs was rapid and sustained, and the growth of >90% of the inoculum was inhibited by day 1. With TMC207 at 8× MIC, less than 0.003% of the initial numbers of RLUs remained by day 2. The reduction in the numbers of RLUs caused by RIF, on the other hand, was much slower: it took 12 days to bring down the number of RLUs by ≥90%, and the efficacy was similar at all the three concentrations. The drugs at 2× MIC ranked the activities SQ109 > RIF TMC207, but the activities of the drugs at 4× and 8 MIC ranked as follows: TMC207 SQ109 > RIF.
In drug combinations, SQ109-TMC207 was extraordinarily potent, with ≥90% killing occurring as early as 1 day after exposure, even in the group exposed to 2× MIC of TMC207. The combination of RIF with either SQ109 or TMC207 did not materially improve the rate of killing over that achieved by either drug by itself. The efficacies of the combinations were SQ109-TMC207 TMC207-RIF > SQ109-RIF.
Assessment of bacterial viability by use of the luciferase reporter assay (RLU estimation) is sensitive, and that result showed a good correlation with the CFU data (17); however, the method may not differentiate static and cidal effects, and the results should be interpreted along with those of other methods to estimate the numbers of viable bacteria.
The Bactec radiometric growth assay measures the release of radioactive carbon due to bacterial respiration and should also be viewed as a measure of metabolic inhibition. In this experiment, one of several similar experiments, the GI for the control inoculum (no drug treatment) at day 1 was 559, so a decrease in the GI to ≤56 was considered to be ≥90% killing of M. tuberculosis H37Rv.
When the individual drugs were used at 2× MICs, none of the drugs caused a ≥90% reduction in the GI by day 8 (Table (Table5).5). In fact, even though the decrease in the GI by SQ109 was dose dependent, the drug at any concentration did not achieve ≥90% killing by day 8. In contrast, TMC207 at >2× MIC achieved a ≥90% reduction in the GI by day 7, and RIF at >2× MIC achieved the same reduction by day 6.
In drug combinations, the activity of the combination of SQ109 and TMC207 at 4× and 8× individual MICs was by far the best, with a ≥90% reduction in the GI occurring on day 3 for both concentrations. The activity of the combination of SQ109 and RIF was not far behind that of SQ109 and TMC207, with a ≥90% reductions in the GI occurring on day 3 (8× MIC) and day 4 (4× MIC). At 4× and 8× MIC, the order of activity was SQ109-TMC207 > SQ109-RIF > TMC207-RIF, with a ≥90% reduction in killing occurring on days 3, 4, and 6, respectively. At higher concentrations, 8× MIC, the order of activity of the drug combinations was SQ109-RIF = SQ109-TMC207 > TMC207-RIF.
Table Table66 shows the log10 numbers of CFU recovered from M. tuberculosis H37Rv treated with SQ109, TMC207, or RIF (alone or in combination) at various times. The M. tuberculosis H37Rv inoculum at the start of the experiment was approximately 107 bacteria, and this number increased in untreated cultures to approximately 1011, a 4-log10 increase, by 2 weeks. To inhibit 90% of the original inoculum or greater, any drug or drug concentration needed to decrease the CFU count to 106 CFU or less.
Alone, SQ109 and RIF at 4× and 8× MICs achieved >90% killing within 4 days, while TMC207 did not. SQ109 alone at 4× and 8× MICs was more effective than the drug combinations. However, the TMC207 and RIF combinations were more effective in killing M. tuberculosis than the single drugs. Interestingly, by 15 days after the start of the experiment, the combination of SQ109 plus either TMC207 or RIF at 8× MIC was exceptionally potent (see the boldface data in Table Table6):6): these drug combinations induced a decrease of >5 log10 CFU from the count in the initial inoculum and a decrease of >9 log10 CFU from the final control CFU count. The order of activity of the individual drugs for reducing the numbers of CFU was SQ109 > RIF > TMC207; the order of activity for the drug combinations varied, depending upon the day on which the data were collected. At 4 days, the order was SQ109-TMC207 = TMC207-RIF > SQ109-RIF; by 15 days, however, the SQ109-containing regimens were clearly better, and the order was SQ109-TMC207 = SQ109-RIF > TMC207-RIF. In fact, however, the actual rates of killing by all the combinations were quite similar.
The PAE determines the time required for the bacteria to recover following a single exposure to a drug below or at its maximum concentration in plasma. Unlike in vitro susceptibility testing, in which the organisms were continuously exposed to the antimicrobial agent, in the pulse-exposure experiments, the bacteria (M. tuberculosis H37Rv) were exposed to drug for only a short time interval. The bacterial suspension was then washed to remove the drug or diluted well below the drug MIC levels, and the organisms were allowed to recover and grow. The mean cumulative GIs of the drug-free control and each drug were plotted against time. Two hours of exposure to SQ109 did not induce any PAE (data not shown), which might be expected for a drug that does not kill by metabolic means but, rather, kills by affecting late stages in cell wall synthesis. In contrast, a 2-h exposure to TMC207 by itself resulted in a PAE of over 9 h. INH, the positive control in this experiment (Bactec radiometric growth assay MIC, 0.06 μg/ml), had a PAE of over 15 h, which has been reported elsewhere (15). SQ109 improved the TMC207 PAE when TMC207 was used in combination with SQ109, and the PAE of this combination was 13.6 h, which is about the same as that of INH.
The outcome of treatment with any drug combination depends on the drug target(s) and the MOAs of the individual agents, as well as the interactions of the drugs with mammalian physiologic processes. An individual drug can enhance (synergistic activity), suppress (antagonistic activity), remain indifferent, or be additive to the activity of the companion drug solely on the basis of its MOA; or it can affect the in vivo efficacy of the companion drug in either a positive or a negative way by modifying host absorption, distribution, metabolism, and excretion (ADME).
Since multidrug anti-TB chemotherapy is the rule, it is essential to determine the interactions among drugs to arrive at a safe and effective regimen(s). Several studies of TB drug combinations with mice point out that even the current standard-of-care regimen for TB may not be optimal, on the basis of pharmacokinetics and outcomes in mice (6). These data suggest that antagonism between INH and the combination RIF-PZA occurs during the initial phase of TB chemotherapy in mice.
The empirical approach to a new anti-TB drug regimen is not appropriate today. A more rational approach to the development of a new anti-TB drug regimen under today's circumstances is to begin assessing new anti-TB drug candidate interactions while clinical trials are ongoing to determine which of the new anti-TB drugs, when they are registered, will work well together in the clinic. In the present in vitro collaboration, all drugs were bactericidal individually, and no combinations were antagonistic, with synergistic or enhanced interactions predominating.
It should be noted that positive (synergistic activity) or negative (antagonistic activity) drug interactions can be evaluated only at concentrations on the exponential portion of the drug response curve for each drug, since once the plateau of killing is reached by a highly potent drug, a significant increase in killing by a second drug would require that the second drug affect a totally separate subset of organisms. At super-MIC levels, such as those that are required in time-kill studies, the synergy or antagonism outcomes may not be so obvious because each of the drugs is operating at its optimal level. What can be assessed is any change that occurs in the amount of time that it takes to kill an individual bacterium when two drugs with different targets are used in combination. Using this definition of synergy, we could assess the combined effects of two drugs on the rate of killing.
Overall, SQ109-RIF was the most effective drug combination in all the assays, followed closely by SQ109-TMC207. Synergy between SQ109 and RIF was previously demonstrated both in vitro (5) and in vivo (12). Similarly, TMC207 was synergistic when it was used with pyrazinamide but not with RIF (8).
With the in vitro data suggesting that SQ109 and TMC207 work well together, we have initiated a series of studies with murine models of TB to confirm and extend these observations and determine which other drugs will complement the synergistic effects of SQ109 and TMC207 in combination.
Published ahead of print on 12 April 2010.