INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Rifampin (R) is the key component of the short-course regimens for the treatment of tuberculosis (1), but it has to be given daily otherwise its activity against Mycobacterium tuberculosis is significantly reduced (2). When the drugs are administered daily or three times weekly, the operational requirements for provision of directly observed treatment (DOT), a key component of the World Health Organization (WHO) Directly Observed Treatment Short-course (DOTS) strategy (3) are significant, particularly in areas where the health infrastructure is poor, and accessibility of services is difficult. On the other hand, if the frequency of drug administration could be reduced to no more than once weekly while maintaining the efficacy of the treatment, the implementation of DOT would be easier.

Rifapentine (P), which has recently been approved by the U.S. Food and Drug Administration for the treatment of pulmonary tuberculosis, offers this possibility. The pharmacokinetic properties of P are very favorable, as its serum peak level (C_max) and half-life (t_1/2) are significantly greater than those of other available rifamycin derivatives (2). Consequently, significant bactericidal activity was still observed in mice treated intermittently with P at a frequency of up to once fortnightly (2, 6). The addition of isoniazid (H), also given intermittently, enhanced the bactericidal activity of P in both immunocompetent (normal) and nude mice (6), suggesting that PH given once weekly may provide an effective intermittent regimen for both the prevention and treatment of tuberculosis.

Further experiments confirmed the expectation that, after 8 wk of treatment, strong bactericidal activity was consistently demonstrated in mice treated with various once-weekly P-containing combined regimens. Although the effect was approximately 1 log₁₀ less bactericidal than that of the standard daily regimen, i.e., 6-times-weekly, RH plus pyrazinamide (Z), the difference only represented a fraction of the overall 4- to 5-log₁₀ magnitude of killing effect (7). The addition of streptomycin (S) improved the bactericidal activity of once-weekly PHZ, and the activity of once-weekly PHZS was further enhanced when it was preceded by an initial 2-wk daily phase including S. In that case, bactericidal effect was similar to that of the standard daily RHZ regimen (7). Furthermore, during the 8-wk course of treatment, all of once-weekly P-containing regimens were able to prevent the selection of R-resistant mutants (7).

The objectives of the current experiment were first to confirm, in an experiment of longer duration, that once-weekly P-containing regimens displaying promising bactericidal activity against M. tuberculosis were able to prevent the selection of R-resistant mutants; and, second, to compare the sterilizing activities of the tested regimens with that of standard short-course regimens (5).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mice

Female Swiss normal mice, 4 wk old, were purchased from the Janvier Breeding center, le Genest Saint-Isle, France.

M. tuberculosis Strain

The virulent H37Rv strain of M. tuberculosis was grown on Löwenstein-Jensen (L-J) medium. Colonies were subcultured in Dubos broth (Diagnostics Pasteur, Paris, France) for 7 d 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 bacillus Calmette-Guérin (BCG), and was further diluted 5-fold for mouse inoculation. The minimum inhibitory concentrations (MICs, in µg/ml) for strain H37Rv were 0.25 for R, 0.06 for P, 0.1 for H, and 2.0 for S on 7H11 agar medium, and 2.0 for ethambutol (E), and 10 for Z on L-J medium (pH 5.5).

Antimicrobial Agents

P, R, and Z were kindly provided by Hoechst Marion Roussel, Romainville, France, and H by Roche, Neuilly, France. S was purchased from Solvay Pharma, Suresnes, France, and E from Wyeth-Lederle, Puteaux, France.

Infection of Mice

A total of 395 mice were intravenously infected with 0.5 ml of a bacterial suspension containing 4.3 × 10⁶ colony-forming units (cfu) of M. tuberculosis H37Rv.

Chemotherapy

As shown in the scheme of the experimental design summarized in Table 1, there were five control groups and seven test groups treated with various once-weekly P-containing regimens. The control groups were uninfected and untreated (A), infected and untreated (B), treated with the standard American Thoracic Society 6-mo regimen of 2 mo of daily (six times a week) HRZ followed by 4 mo of daily HR (C), treated with the standard WHO 8-mo regimen of 2 mo daily HRZ followed by 6 mo of daily HE (D), or treated with 6 mo of once-weekly P (E). Both Groups B and E were negative controls that served to define the magnitude of the progressive bacillary infection (Group B) and to determine if monotherapy with rifapentine would select rifampin-resistant mutants (Group E). Treatment began on Day 14 after infection, designated as D0, and continued for 6 to 8 mo to mimic the usual duration of treatment for human pulmonary tuberculosis. All of the drugs, except S which was given by subcutaneous injection, were suspended in 0.05% agar-containing distilled water at the desired concentrations and were administered through an esophageal cannula (gavage); the suspensions were prepared weekly and stored at 4° C. The drugs were given at the following dosages in mg/kg: R 10, P 10, H 25, and 75 when it was given, respectively, at a frequency of six times weekly (daily) and once weekly, Z 150 (daily) and 300 (once weekly), S 200, and E 100. The dosages were the same as those in our previous experiments (2, 6), and were selected to be equipotent with those achieved in humans (2, 12). In addition to the dosage of 10 mg/kg of P, Groups I and K were treated, respectively, with P at dosages of 5 and 15 mg/kg, to evaluate the impact of different dosages of P on the bactericidal effects.

Assessment of Infection and Treatment

The severity of infection and the effectiveness of the treatment were assessed by comparing the survival rate, spleen weights, gross lung lesions (the severity of which were scored from 0 to ++, the latter referring to a lung that was extensively occupied by tubercles), and total cfu counts in the spleens and lungs between the groups (2, 6, 7). In Group B, 15, 30, and 15 mice were planned to be killed, respectively, on 1, 14, and 28 d after infection, or D13, D0, and 2 wk in relation to the date when treatment began, to provide the baseline values. At D0 and after 2 wk, three consecutive 10-fold diluted suspensions were plated under the volume of 0.2 ml per tube onto three tubes of plain and three tubes of 40 µg/ml R-containing L-J medium per dilution. However, after 6 and 8 mo of treatment, because very few viable organisms were expected in the organs, the suspension from individual organ was plated without dilution on 10 tubes of plain and five tubes of R-containing L-J medium. Results of the cultures were recorded after incubation at 37° C for 6 wk.

The lower limit of detectability of the cfu per organ was either 0.77 or 0.24 log₁₀, which refers respectively to a single colony detected on one of the three tubes with a 10-fold dilution or on one of the 10 tubes plated with undiluted suspension. A bactericidal effect of the treatment was defined as a statistically significant decrease in the mean number of cfu in the treated mice from the pretreatment value.

After 6 or 8 mo of treatment, whenever the culture was positive on plain L-J medium with or without R-resistant mutants isolated on R-containing L-J medium, the susceptibility to R and H for all organisms isolated on plain L-J medium was tested by the proportion method (13). In brief, the colonies were collected and subcultured in Dubos broth for 7 d at 37° C, and the turbidity of the resulting suspension was adjusted to 1 mg/ml; serial 10¹, 10³, and 10⁵ dilutions were prepared and plated on drug-free L-J medium and the medium containing R at the concentration of 40 µg/ml or H at 0.1, 0.2, 1.0, and 10.0 µg/ml. The growth was read and recorded after 42 d of incubation at 37° C. A drug-resistant strain was defined as a strain containing at least 1% of the cfu multiplying on L-J medium containing 40 µg/ml of R (14) or 0.2 µg/ml of H (13), and the resistance rate was defined as the number of mice from which a M. tuberculosis strain resistant to R or H was isolated out of the number of mice examined. Only those cultures, that yielded R-resistant mutants at isolation, were confirmed R-resistant at secondary drug susceptibility testing.

Because a significant number of mice were culture-negative by the end of 6 or 8 mo of treatment, the effectiveness of the treatment regimens was based on the culture positivity rate, the resistance rate, and the treatment failure rate. Treatment failure refers to a mouse whose spleen and/or lung were culture-positive, regardless of the number of cfu, by the end of treatment (8 mo in the current experiment except for mice of Groups C and E), or in which drug-resistant strains had emerged after 6 mo of treatment. The treatment failure rate was based on the number of mice that fulfilled the previously mentioned definitions.

Statistical Analysis

The cfu data presented heterogeneous variances. Therefore, in order to make multiple comparisons of the treated groups of mice, we adopted a nonparametric analysis according to the following strategy: pairwise comparisons were made on ranked data using the Wilcoxon-Mann-Whitney test (15), but the critical value for a type-1 error was corrected in order to account for the six comparisons that were performed. Therefore, (1  a_corrected) was equal to (1  a)⁶, and accordingly the critical value for a significant difference with a type-1 error of 5% was 2.632 instead of 1.96.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Survival Rate

As expected, the survival rate of the uninfected, untreated mice (Group A) was 100% whereas, after intravenous infection with more than 10⁶ cfu of the virulent strain H37Rv, mice in the untreated controls (Group B) began to die from Day 18 after infection; 95% of them died by Day 28, and all died by Day 36 (Figure 1). Survival of mice in Group E remained at 100%, indicating that the first once-weekly dose of P₁₀ alone could effectively prevent the early deaths caused by infection with M. tuberculosis.

View larger version (20K): [in this window] [in a new window]   Figure 1.   Survival rates of mice during 42 d after intravenous infection with 4.3 × 106 cfu of M. tuberculosis H37Rv. Group B was negative control, i.e., infected but untreated. Treatment began 14 d after infection, and Groups C-L were treated with various regimens specified in Table 1. However, during the period between 14 and 42 d after infection, the treatments for Groups C, D, and F were identical, i.e., RHZ 6/7; and those for Groups I, J, K, and L were also very similar, beginning with 2 wk (RHZ 6/7), followed by once-weekly treatment with combinations of various dosages of P plus H and Z; Groups G and H began with 2 wk P10HZ 6/7, and continued with P10HZ 1/7, the only difference between the two groups was with or without S; and Group E was treated with P10 alone 1/7.

In all treated groups, the survival rates were very high, virtually 100% during the first months of treatment. However, the pattern of the survival curve in Group G was quite different from the rest of treated groups such that 1, 3, and 1 mice died, respectively, after the first, second, and third dose of treatment, but no further deaths were observed despite continuation of treatment with the same regimen at the frequency of 6 d a week or once weekly.

Mean Spleen Weght

The mean spleen weight of the infected mice significantly increased, more than 4-fold, during the initial 14 d after infection. After 2 wk of treatment, among the treated groups only the spleen weights from Groups G, H, and I were available for comparison: the mean value for each of the three groups was significantly smaller than that for control mice killed on D0 of treatment but did not differ significantly among one another. Furthermore, the mean spleen weight did not differ significantly among the treated groups killed after 6 or 8 mo of treatment; all were significantly smaller than the pretreatment value, but none of them returned completely to the baseline level of Group A. These results indicate that all treatments were equally effective in reversing the splenomegaly caused by tuberculosis infection.

Gross Lung Lesions

By the time the treatment began, i.e., D0 or 14 d after infection, severe gross lung lesions (++) were observed in 29 of 30 mice in group B. After 2 wk of treatment, (++) lesions were still observed in all sacrificed mice of Groups G, H, and I. However, after 6 mo of treatment, (++) lesions were observed only in very few (0/15 or 1/15) treated mice, except for Group E in which 6 of 13 mice still had such lesions, apparently because of the emergence of R-resistance. After 8 mo of treatment, (++) lesions were no longer observed in the treated groups. These results indicate that the lung lesions caused by tuberculosis infection were progressively cured by the treatment of all of the combined regimens but not by P monotherapy.

Enumeration of cfu (log₁₀) and Drug Resistance in the Spleens and Lungs

The day after intravenous infection, the mean number of cfu in the spleens, 4.56 ± 0.82 log₁₀, was similar to that in the lungs, 4.40 ± 0.29 log₁₀; within the following 13 d, the cfu increased rapidly, about 2 log₁₀ in the spleens and 3 log₁₀ in the lungs to reach 6.51 ± 0.39 log₁₀ and 7.68 ± 0.44 log₁₀, respectively. Therefore, by the time the treatment began, the mean number of cfu in the lungs was 1 log₁₀ greater than that in the spleens.

As shown in Table 2, after 2 wk of treatment the mean numbers of cfu in Groups G, H, and I were significantly smaller (p < 10⁵) than the pretreatment values both in lungs and in spleens, indicating that such a brief duration of treatment with any of the three regimens exerted substantial bactericidal activity, with at least 2 log₁₀ of killing effects against M. tuberculosis. The reduction of cfu in the lungs did not differ significantly between Groups H and I (p = 0.56), whereas the reduction was significantly greater in Group G compared with Groups H and I (0.76 and 1.09 log₁₀; p = 0.002 and p = 0.0003, respectively). When comparing the cfu counts in the spleen of Groups G, H and I, only counts in Groups G and I were significantly different (p = 0.016). Although Groups G and I had closer mean values than Groups H and I, the significant difference between Groups G and I was due to the small variation into the cfu counts in Group G. At the end of 6 mo of treatment and onward, the mean numbers of cfu in all treated groups were significantly lower than the pretreatment values.

As expected, at 6 mo the organs of Group C mice were all culture-negative, but those in two of the 15 Group D mice remained positive; however, at 8 mo, the cultures of all 14 mice of Group D were negative. In Group E mice, which received P monotherapy, all surviving mice were culture-positive at 6 mo with a mean 5.41 ± 0.77 log₁₀ cfu in the lungs and 4.32 ± 1.14 log₁₀ cfu in the spleen, and R-resistant strains were isolated from all (Table 3). The organs of Group F mice, treated with 2 mo of daily RHZ followed by 4 mo of once-weekly PH, were also culture-negative at 6 mo and remained negative at 8 mo.

Most of the six Groups G to L regimens, which consisted of 2 wk of 6 d a week treatment, followed by 7.5 mo of once-weekly treatment, did not give favorable results. At 6 mo, only the organs from mice of Groups G and K were all culture-negative. At 8 mo a positive culture with organisms resistant to R was obtained from the lungs of a mouse of Group K, whereas all the cultures from Group G mice remained negative (Tables 2 and 3). Mice of Group H had positive cultures at 6 or 8 mo, and two of the four positive cultures at 6 mo yielded organisms resistant to R (Table 3). The only difference between the regimens for Groups G and H was that the former regimen was supplemented with S. The regimen for Group J was similar to that for Group F, except the intensive daily phase with HRZ was 2 wk in Group J and 2 mo in Group F. As shown in Table 3, at 6 mo the culture-positive rate in Group J mice was 3 of 15, and one of the three isolates was resistant to R whereas all Group F mice were culture-negative from 6 mo onwards.

The regimens for Groups I, J, and K did not give favorable results despite an increase in the dosage of P from 5 to 10 and 15 mg/kg, respectively. However, as shown in Table 3, cultures were more frequently positive in Group I than in the other two groups, and Group I mice were the only mice in the current experiment from which H-resistant mutants were isolated, i.e., in 2 of the 6 mice and in 1 of 3 mice that were culture-positive at 6 and 8 mo, respectively. As a treatment failure with R-resistance was observed in each of the Groups J and K at 6 or 8 mo, the increased sterilizing activity resulting from the increased P dosage was not sufficient to overcome the risk of selecting R-resistant mutants.

Finally, regimen L was the only one among those not supplemented with S to give favorable results at 8 mo without any selection of drug-resistant strains during the course of treatment. The only difference between regimen L and regimen J was that the former was supplemented with 6 d a week H throughout the whole duration of treatment. That supplement was sufficient to prevent the selection of R-resistant mutants.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

While the daily short-course regimen with R throughout the whole duration of treatment (Group C) was able to render all organs of mice culture-negative at 6 mo, the daily short-course regimen with R only during the initial 2-mo intensive phase was unable to do so at 6 mo and required 8 mo to give similar results. Thus, the justification for treating pulmonary tuberculosis patients with a 6-mo or 8-mo regimen according to the duration of R administration (5) is experimentally confirmed.

In terms of preventing early deaths, it is interesting that monotherapy with once-weekly P (Group E) was so effective. However, treatment failures resulting from the emergence of R-resistant mutants invariably occurred when the treatment continued up to 6 mo.

Despite the half-life of P being close to 24 h and that of R at least three times shorter (2), 10 mg/kg P and 10 mg/kg R given 6 d a week during the initial 2-wk phase (Group H) were similarly effective at the end of the initial phase and did not confer subsequent benefit. As the daily administration of P might be associated with a higher risk of toxicity due to drug accumulation, and the experimental activity of daily rifapentine was not significantly better than that of daily rifampin, there is no apparent benefit of giving daily P for the treatment of human tuberculosis, even only during a short initial period.

A previous experiment, in a murine tuberculosis model similar to the current one, demonstrated that the 2-mo initial intensive phase of treatment with daily RHZ killed at least 4 to 5 log₁₀ of viable organisms, respectively in the spleen and lungs. Consequently, the residual microbial population was no more than 2 log₁₀ per organ (7), and likely no longer contains any drug-resistant mutants. After the 2-mo intensive phase with RHZ, a 4-mo continuation phase with once-weekly PH (Group F) was equally effective as with daily RH (Group C) and more effective than a 4-mo continuation phase with daily HE, as a mouse treated with the latter regimen was still culture-positive at 6 mo. That finding suggests that once-weekly PH in the continuation phase of tuberculosis therapy is as effective a sterilizing regimen as daily RH and better than daily HE, providing that the bacterial population at the end of the intensive phase is low. In one clinical trial which evaluated once-weekly rifapentine plus isoniazid in the continuation phase, the rate of relapse after treatment was increased among patients who were still sputum culture-positive at the end of the intensive phase of therapy (16). For patients who have a good bacteriologic response during the initial phase of treatment, rifapentine, which was recently approved for use in the United States, now provides for once-weekly dosing during the last 4 mo of therapy. This should simplify the treatment of pulmonary tuberculosis, facilitate its supervision, and reduce its cost.

However, the 2-mo initial phase of daily treatment with RHZ still poses significant operational difficulties for implementation of DOTS, and one of the possible solutions is to reduce the daily phase from 2 mo to 2 wk before beginning once-weekly administration of a P-containing regimen (7). The cfu counts in mice of Group I demonstrated that at the end of 2 wk of intensive daily phase with RHZ, only 2 log₁₀ of viable organisms per organ had been killed, so that the bacterial load in the lungs remained relatively large, above 5 log₁₀ of drug-susceptible organisms and possibly included mutants resistant to R or H. The treatment failures caused by the emergence of R-resistance that were observed at the end of 6 or 8 mo in a few mice of Groups H, J, and K, likely reflect the inability of the once-weekly PHZ to prevent in all cases the selection of drug-resistant mutants still present in the large bacterial population remaining after 2 wk of intensive treatment with daily RHZ. Therefore, in spite of its good sterilizing activity, the once-weekly PHZ combination should be enhanced to cope with the large microbial populations that remain when the duration of the initial daily phase is limited to 2 wk. This may be especially important for patients with human immunodeficiency virus (HIV) infection who appear to be more likely to develop acquired rifampin-monoresistant disease when treated with once weekly rifapentine-isoniazid (17).

A possibility to enhance the ability of the once-weekly PHZ to prevent the selection of R-resistant mutants was to increase the P dose. To test that hypothesis, mice of Groups I, J, and K were given, respectively, 5, 10, and 15 mg/kg of P during the 7.5 mo of the once-weekly drug combined continuation phase. The sterilizing activity of P-containing regimens was clearly enhanced when the normal dose of 10 mg/kg was increased to 15 mg/kg, but this higher dose was not able to prevent the selection of R-resistant mutants in all treated mice. This finding suggests that the selection of the R-resistant mutants was related to the size of the residual microbial population and to the insufficient bactericidal activity of the drugs combined with P, namely H and Z. However, once-weekly combination HZ is not totally deprived of bactericidal activity because, when the P dosage was reduced to 5 mg/kg, the overall sterilizing activity of the regimen was strongly reduced but H-resistant mutants were selected in 3 of 27 mice, indicating that once-weekly H had some degree of bactericidal activity. Nevertheless, once-weekly combination HZ suggests that Z was not been able to prevent the selection of H-resistant mutants and that its contribution in the once-weekly regimen was limited or absent.

Another possibility to enhance the ability of the once-weekly PHZ to prevent the selection of R-resistant mutants was to add a fourth drug, such as S. Among the six groups of mice that received only 2 wk of initial daily phase followed by 7.5 mo of once-weekly continuation phase, only Group G mice, in which the treatment was supplemented with daily S during the initial phase and once-weekly S during the continuation phase, were all culture-negative at 6 mo and remained so at 8 mo. At the end of 2 wk of treatment, the cfu counts in the lungs of Group G mice were below 5 log₁₀, i.e., about 1 log₁₀ lower than those in other groups, and likely contained no mutants resistant to R or H. The promising results from group G suggested that the addition of S enhanced not only the overall bactericidal activity but also the ability to prevent the treatment failures resulting from the selection of resistant mutants. The findings were in agreement with our previous results (7) but do not permit one to determine whether the role of S was decisive during the initial daily phase, the once-weekly continuation phase, or both. Further investigations are needed to address this question.

The last possibility to enhance the ability of the once-weekly PHZ to prevent the selection of R-resistant mutants was to give H daily. Programatically, this would be similar to use of self-administered dapsone/clofazimine in the treatment of leprosy. The regimens of multidrug therapy (MDT) for leprosy consist mainly of R given monthly under supervision and daily self-administered dapsone with or without clofazimine (18). Because more than 10 million leprosy patients have been cured by these regimens with virtually no emergence of R-resistance (18), it was of interest to evaluate in tuberculosis the additional impact of daily H upon the bactericidal effect of 7.5 mo of once-weekly PHZ after 2 wk of daily phase with RHZ. Among the mice of Group L treated with that regimen, only one was culture-positive at 6 mo without R-resistant mutants and all were culture-negative at 8 mo, suggesting that increasing the frequency of H administration might compensate for the short duration of initial phase as well as the lack of the S supplement.

Finally, the five deaths that occurred only during the initial 3 d of treatment among mice in Group G were of concern. Because the major difference between regimen G and others was that the former contained S, which was injected 6 d a week during the initial 2-wk phase (i.e., between Day 14 and Day 28 after infection), it is likely that the early deaths in Group G were related to the injection of S. A similar phenomenon was also observed in our previous experiment in which approximately 40% of mice died exclusively after the first two injections of S (7). At that time, we concluded that the deaths were most likely due to an accident caused by the injection of S rather than by the toxicity of S (7). However, the mortality of mice treated with daily injections of S at the same 200 mg/kg dose had never been a problem in our earlier experiments (9) and if the deaths in the current experiment had resulted from S toxicity, one would have expected more deaths to occur during the subsequent treatment in both experiments. Further toxicologic investigations are in progress.

In conclusion, once-weekly P-containing combined regimens were effective only if the microbial population in the organs had been greatly reduced by a daily intensive initial phase of sufficient duration or if the bactericidal activity of the once-weekly phase was enhanced by the addition of S or the administration of daily H. However the optimal duration of once-weekly S remains uncertain, and could possibly be much shorter than the full 7.5 mo used in regimen G. It is also possible that the addition of another bactericidal drug, such as one of the new fluoroquinolone compounds (19, 20), might provide for a highly effective, once-weekly fully oral regimen.