Comparison of a Novel Regimen of Rifapentine, Tedizolid, and Minocycline with Standard Regimens for Treatment of Pulmonary Mycobacterium kansasii

Effects of static drug concentrations.

The MICs of the drugs used in the study, based on CLSI methodology but using the Middlebrook 7H9 with 10% oleic acid-albumin-dextrose-catalase (OADC) (herein termed broth), are shown in Table 1. M. kansasii on day 4 of log-phase growth in broth was coincubated with 7 different concentrations of either rifapentine, tedizolid, and minocycline for 7 days, in order to determine the efficacy (maximal kill [E_max]) of each drug against axenic cultures of M. kansasii. The results of the inhibitory sigmoid E_max model are shown in Fig. 1. The E_max shown in figure is in comparison to the nontreated control on day 7 and also demonstrates the kill below day 0 bacterial burden or stasis. The tedizolid kill below stasis was 4.22 ± 0.11 log₁₀ CFU/ml, that for rifapentine was 5.40 ± 0.11 log₁₀ CFU/ml, and that for minocycline was 5.01 ± 0.13 log₁₀ CFU/ml. This means that tedizolid, rifapentine, and minocycline have killing efficacy against M. kansasii and should be tested further as a three-drug combination mimicking human-like pharmacokinetics and compared with the current standard regimens for efficacy.

HFS-Mkn study comparing standard and experimental regimens.

We used the intracellular HFS-Mkn model (15, 16) to compare the killing efficacy of tedizolid, rifapentine, and minocycline as a two- or three-drug combination with the ATS or the BTS regimen. There were 18 HFS-Mkn units treated with tedizolid, rifapentine, minocycline, and clarithromycin as monotherapy as well as two- and three-drug combinations, with two replicates per regimen, including the nontreated controls. Figure 2A shows the drug concentration-time profile achieved in the HFS-Mkn by sampling the central compartment at six time points over 24 h. The measured drug concentrations, alone or in combination, were used to calculate the drug exposure (area under the concentration-time curve from 0 to 24 h [AUC_0–24]/MIC), as shown in Fig. 2B. The number of viable THP-1 cells in the HFS-Mkn treated with different regimens is shown in Fig. S1 in the supplemental material.

The mycobacterial growth indicator tube (MGIT)-derived time to positivity (TTP) in the HFS-Mkn treated with either monotherapy or combination therapy at each sampling day is shown in Fig. 3A. The longer the TTP, the lower the intracellular bacterial burden in the HFS-Mkn. The TTP of the inoculum (stasis) was 4.82 days. On day 28, as shown in Fig. 3A, the TTP of the HFS-Mkn treated with tedizolid (1.96 days) or minocycline (3.13 days) was longer than for the nontreated controls (1.51 days); however, it failed to keep the bacterial burden below stasis. In contrast, in the HFS-Mkn treated with rifapentine monotherapy, the TTP increased to 12.13 days on study day 21 and afterward the trajectory changed, indicating failure of rifapentine monotherapy, as the TTP at the end of study was 9.21 days. On day 28, the TTP of the BTS regimen (13.67 days) was significantly different from that of the ATS regimen (10.67 days; P = 0.002). The rifapentine-tedizolid two-drug combination showed a TTP (13.33 days) similar to that of the BTS regimen and was not statistically different (P > 0.19). The longest TTP was recorded with the rifapentine-tedizolid-minocycline combination, 14.98 days. However, none of the treatment regimens showed total microbial kill or sterilization of HFS-Mkn by the TTP measure.

Figure 3B shows the bacterial kill slopes of the HFS-Mkn with different treatment regimens, based on the readouts for log₁₀ CFU per milliliter. In Fig. 3B, similar to the TTP results, tedizolid and minocycline failed to keep the bacterial burden below stasis (4.07 log₁₀ CFU/ml). Clarithromycin maximal kill was 0.71 log₁₀ CFU/ml on day 7, and clarithromycin monotherapy failed after day 21. In comparison, the maximal kill with rifapentine monotherapy was 2.20 log₁₀ CFU/ml on day 14; afterward, the direction of the slope changed, showing therapy failure. However, the bacterial burden in the rifapentine-treated HFS-Mkn was still below stasis on day 28. The maximal kill with the ATS regimen was 2.25 log₁₀ CFU/ml on study day 7, after which the regimen failed. In comparison, the BTS regimen continued to kill M. kansasii and the bacterial burden was 1.00 log₁₀ CFU/ml on day 28. The most impressive results were obtained with the rifapentine-tedizolid-minocycline regimen (Fig. 3B), which killed the total M. kansasii population in 28 days. However, since the MGIT tubes showed positive results (Fig. 3A), this indicates that the bacterial burden was simply below the limit of detection of the agar method of quantitation (i.e., 0.69 log₁₀ CFU/ml). The overall kill rate of the ATS regimen was −0.038 ± 0.0.038 log₁₀ CFU/ml/day, that of the BTS regimen was −0.083 ± 0.035 log₁₀ CFU/ml/day, and that of the rifapentine-tedizolid-minocycline combination was −0.119 ± 0.031 log₁₀ CFU/ml/day. The kill slope for the rifapentine-tedizolid-minocycline regimen was significantly different from that of the ATS regimen (P < 0.03) but not from that of the macrolide-containing BTS regimen (P = 0.05).

Bacterial strain, cell line, drugs, and other supplies.

We used the M. kansasii standard laboratory strain ATCC 12478 in all experiments. The medium for the culture of the bacteria was either Middlebrook 7H9 broth or Middlebrook 7H10 agar supplemented with 10% oleic acid-albumin-catalase-dextrose (OADC) with culture conditions as reported previously.(15, 16). The human monocyte-derived THP-1 cell line (ATCC TIB-202) cultured in RPMI 1640 medium supplemented with 10% or 2% fetal bovine serum (FBS) was used in the intracellular M. kansasii experiments, including the HFS-Mkn studies (15, 16). Hollow-fiber cartridges were purchased from FiberCell Systems Inc. (MD). Isoniazid, rifampin, rifapentine, minocycline, clarithromycin, and ethambutol were purchased from Sigma-Aldrich (St. Louis, MO), and tedizolid was synthesized by BOC Sciences (NY). The mycobacterial growth indicator tube (MGIT) liquid culture system and MGIT tubes were purchased from Becton, Dickinson, USA.

Determination of MIC and concentration-response studies.

The MIC of the drugs was determined using the broth microdilution method (24), in which Middlebrook 7H9 with 10% OADC was used as the culture medium. The concentration range of the drugs used in the MIC experiments is shown in Table 1. Briefly, M. kansasii cultures were grown for 4 days to obtain the logarithmic-phase culture and the inoculum was prepared to get a bacterial density of 1.5 × 10⁵ CFU/ml. The experiment was set in 96-well plates in triplicate for each concentration, including the nontreated control. The cultures were incubated for 7 days, and plates were inspected for presence of bacterial pellet. The drug concentration that completely inhibited the M. kansasii growth (absence of bacterial pellet) was recorded as the MIC. Experiments were performed twice.

For the rifapentine, tedizolid, and minocycline concentration-response study, the culture condition, inoculum preparation, and concentration range were the same as described above, except that the experiment was set in 15-ml screw-cap tubes with a total volume of 5 ml. M. kansasii cultures were coincubated with the different drug concentrations for 7 days at 37°C under shaking conditions. On day 7, the culture was washed with normal saline to remove carryover drug, serially 10-fold diluted, and inoculated on Middlebrook 7H10 agar supplemented with 10% OADC. The M. kansasii colonies on the agar plates were counted after 10 to 14 days of incubation at 37°C.

HFS-Mkn studies to compare the standard and new drug combination regimens and therapy duration.

A detailed description of the HFS-Mkn has been published elsewhere (16, 25). Briefly, human monocyte-derived THP-1 monocytes were infected overnight with M. kansasii at a multiplicity of infection (MOI) of 1:1. After washing twice with warm RPMI 1640 medium to remove the extracellular bacteria, 20 ml of infected THP-1 cells was inoculated into the peripheral compartment of each HFS-Mkn. In the monotherapy treatment arms, tedizolid was tested at a 200-mg daily dose, whereas rifapentine (80% effective concentration [EC₈₀]) and minocycline (EC₈₀) were selected from the above-mentioned experiment. Clarithromycin was used at a human dose equivalent to 500 mg daily. In the two-drug combination, tedizolid at 200 mg was added to the rifapentine EC₈₀. The drug dose in the three-drug novel combination were tedizolid at 200 mg and rifapentine and minocycline both at EC₈₀. The ATS regimen was a combination of rifampin at 600 mg/day, isoniazid at 300 mg/day, and ethambutol at 15/mg/kg. The BTS regimen was a combination of rifampin at 600 mg/day, ethambutol at 15 mg/kg/day, and clarithromycin at 500 mg/day. We mimicked a 3-h half-life for isoniazid and rifampin, whereas a 12-h half-life was mimicked for ethambutol, tedizolid, rifapentine, and minocycline (16, 25, 26). Drugs were administered into the central compartment via computer-controlled syringe pump over a 1-h period for each drug in the combination. The central compartments were sampled 6 times over a 24-h period, and drug concentrations were measured to validate the concentration-time profile of each drug in the combination and calculation of the drug exposure achieved in the HFS-Mkn. We sampled the peripheral compartment of each HFS-Mkn on days 0, 3, 7, 10, 14, 21, and 28 to determine the number of viable THP-1 cells in each system. The samples were then lysed in phosphate-buffered saline with 0.025% Tween 20, serially 10-fold diluted, and inoculated on agar for estimation of total bacterial burden. A portion of the processed sample was also inoculated in the MGIT tubes as a second method to estimate the bacterial burden in each HFS-Mkn by recording the TTP.

Software and statistical analysis.

ADAPT II was used to perform the pharmacokinetic analysis (27). The measured concentrations were used to calculate the maximum concentration (C_max), AUC_0–24, C_max/MIC, and AUC_0–24/MIC. GraphPad Prism version 7.0 was used to run the inhibitory sigmoid E_max model to determine the relationship between the drug concentration and the bacterial burden (GraphPad Software, CA). Linear regression was used to calculate the kill rate with the experimental regimens and the standard ATS and BTS regimens.