Using Dynamic Oral Dosing of Rifapentine and Rifabutin to Simulate Exposure Profiles of Long-Acting Formulations in a Mouse Model of Tuberculosis Preventive Therapy

ABSTRACT Administration of tuberculosis preventive therapy (TPT) to individuals with latent tuberculosis infection is an important facet of global tuberculosis control. The use of long-acting injectable (LAI) drug formulations may simplify and shorten regimens for this indication. Rifapentine and rifabutin have antituberculosis activity and physiochemical properties suitable for LAI formulation, but there are limited data available for determining the target exposure profiles required for efficacy in TPT regimens. The objective of this study was to determine exposure-activity profiles of rifapentine and rifabutin to inform development of LAI formulations for TPT. We used a validated paucibacillary mouse model of TPT in combination with dynamic oral dosing of both drugs to simulate and understand exposure-activity relationships to inform posology for future LAI formulations. This work identified several LAI-like exposure profiles of rifapentine and rifabutin that, if achieved by LAI formulations, could be efficacious as TPT regimens and thus can serve as experimentally determined targets for novel LAI formulations of these drugs. We present novel methodology to understand the exposure-response relationship and inform the value proposition for investment in development of LAI formulations that have utility beyond latent tuberculosis infection.

. Dose linearity of rifapentine and rifabutin in uninfected female BALB/c mice. • Fig. S2. Dosing simulations for rifapentine and rifabutin to achieve indicated target plasma concentrations in the first PK/PD study. • Fig. S3. Simulated and observed rifapentine plasma concentrations and PK parameters from the first PK/PD study. • Fig. S4. PK/PD relationships of rifapentine and rifabutin in the first PK/PD study. • Fig. S5. Simulated and observed rifabutin plasma concentrations and PK parameters from the first PK/PD study. • Fig. S6. Exposure profiles and dosing simulations for rifapentine in the second PK/PD study. • Fig. S7. Rifabutin exposure profiles and dosing simulations, and observed exposures in the second PK/PD study.
• Table S2. PK parameters associated with oral dosing of rifapentine and rifabutin in uninfected male BALB/c mice.
• Table S3. Original and final experiment schemes for the first PK/PD study evaluating rifapentine and rifabutin regimens designed to maintain stable, predefined target plasma drug concentrations. • Table S4. Summary of M. tuberculosis H37Rv lung CFU counts in first PK/PD study.
• Table S5. Oral dosing regimens designed to simulate LAI exposures of rifapentine and rifabutin in the second PK/PD study. • Table S6. Experiment scheme for the second PK/PD study evaluating simulated LAI regimens of orally dosed rifapentine and rifabutin. • Table S7. Summary of M. tuberculosis H37Rv lung CFU counts in second PK/PD study.
• Table S8. Description and CFU data for bacterial suspensions used for aerosol infections of mice.
• Table S9. Summary of M. bovis rBCG30 lung CFU counts in first PK/PD study. • Table S10. Summary of M. bovis rBCG30 lung CFU counts in the second PK/PD study.

Data Files
• Data File S1. Individual mouse PK data from PK study in female and male BALB/c mice. o S1.1. Individual mouse PK data from rifapentine-dosed female BALB/c mice. o S1.2. Individual mouse PK data from rifabutin-dosed female BALB/c mice. o S1.3. Individual mouse PK data from rifapentine-dosed male BALB/c mice. o S1.4. Individual mouse PK data from rifabutin-dosed male BALB/c mice • Data File S2. Individual mouse CFU and PK data from the first PK/PD study.  For rifapentine, regimens were designed to maintain trough plasma concentrations at 0.18, 0.6, 2, and 3.5 µg/mL (panels A-D, respectively). Each rifapentine regimen included a single loading dose, which was the first dose administered when treatment started on Day 0. The Day 10 PK data for rifapentine were not yet available and thus were not incorporated into the PK model used for these simulations. For rifabutin, regimens were designed to maintain trough plasma concentrations at 0.045 µg/mL and 0.15 µg/mL (panels E and F, respectively). Due to the high peak-to-trough predicted plasma exposures, regimens were also designed to maintain average plasma concentrations at 0.045 µg/mL and 0.15 µg/mL (panels G and H, respectively). In all panels, the target plasma concentration is indicated in red, and the simulated drug exposures are indicated in blue. For BID (bis in die, twice daily) dosing, doses were administered 10 and 14 hours apart (7am and 5 pm daily)

Fig. S3
. Simulated and observed rifapentine plasma concentrations and PK parameters from the first PK/PD study. Data for regimens with trough plasma target concentrations of 0.18, 0.6, 2, and 3.5 µg/mL are presented in Panels A-D, respectively. The red line indicates the plasma target concentration. Open circles represent the observed plasma concentrations from individual samples. The blue line represents the simulated plasma exposures (based on simulations from model fittings to pilot PK data in uninfected mice, including Day 10 PK data), and the yellow line represents the model fitting to the observed plasma exposures. The lower limit of quantification for rifapentine was 0.10 μg/mL. Parameter estimates for apparent clearance (CL/F), apparent volume of distribution (V/F), and the absorption rate constant (Ka), and the associated percent relative standard error (%RSE) for each regimen are provided below each graph. All individual mouse PK data are provided in Data File S2.    Table 4, with detailed descriptions of the dosing provided in Table S5. Exposure profiles for a single simulated LAI dose administered at Day 0 are shown in Panels A, C, and E. Exposure profiles for two simulated LAI doses administered at Day 0 and at Week 4 are shown in Panels B and D.

Fig. S7. Rifabutin exposure profiles and dosing simulations (A-B), and observed exposures (C-D) in the second PK/PD study.
In all panels, the red line indicates the plasma target concentration, and the blue line represents the predicted plasma exposures for the orally-dosed regimen designed to achieve the desired exposure profile. In Panels A and B, yellow shading indicates the desired exposure profile of a simulated LAI dose of rifabutin. Each exposure profile represents a single simulated LAI dose administered at Day 0. An overview of each regimen is presented in Table 4, with detailed descriptions of the dosing provided in Table S5.
In Panels C and D, the open circles represent the observed plasma concentrations, and the yellow lines represent the modeled exposures based on fitting to the observed data. Modeling of rifabutin exposures was limited because rifabutin was often not detected in mouse plasma at the 9 and 24 hour sampling time points. The lower limit of quantification for rifabutin was 0.05 μg/mL. All individual mouse PK data are provided in Data File S3.       *120 mice were ordered and immunized with rBCG30; the 5 extra mice were allocated as follows in the final study scheme: 3 mice sacrificed at the time of second M. tuberculosis challenge infection (Week -6), annotated as previously infected mice; and 1 mouse was added to each of the lowest RPT and RFB dosing regimens. **Five new, previously uninfected mice were included in the second M. tuberculosis challenge infection and were sacrificed the day after infection to determine the implantation associated specifically with the second infection. ***One mouse died during blood sampling on Day 0; therefore, CFU data are only available for 4 mice in this group.  Table S4. Summary of M. tuberculosis H37Rv lung CFU counts in first PK/PD study. "Week 3 (all)" represents the combination of data from the Week 3 and Week 3 + 1.5 days time points. Data represent the mean (standard deviation) for each regimen at each time point. Unless otherwise noted, n = 5 mice per group per time point. See study scheme in Table S3. Individual mouse CFU data are presented in Data S2.

Regimen description
Oral dosing M. tuberculosis H37Rv mean (SD) log 10 CFU/lung at the following time points: During treatment Table S5. Oral dosing regimens designed to simulate LAI exposures of rifapentine and rifabutin in the second PK/PD study. Regimens were administered for a total of 8 weeks, which was divided into 14 four-day dosing periods. Gray and blue shading indicate dosing during Weeks 1-4 and Weeks 5-8 of the study, respectively. For each simulated LAI regimen, the indicated dose was administered twice daily. For the 1HP positive control regimen (not in table), rifapentine at 10 mg/kg was dosed 1 hour before isoniazid at 10 mg/kg. All regimens were administered 7 days/week.

Dosing period
Oral rifapentine dose (mg/kg) for each of the following regimens: Oral rifabutin dose (mg/kg) for each of the following regimens: Table S6. Experiment scheme for the second PK/PD study evaluating simulated LAI regimens of orally dosed rifapentine and rifabutin. The regimens are summarized in Table 4, with detailed descriptions of the dosing provided in Table S5.
*Includes 8 mice that were not previously infected with rBCG30.

Regimen description
M. tuberculosis H37Rv mean (SD) log 10 CFU/lung at the following time points: During treatment Table S8. Description and CFU data for bacterial suspensions used for aerosol infections of mice. Samples were cultured on selective 7H11 agar without selective drugs (plain agar) and selective 7H11 agar supplemented with 40 µg/mL hygromycin (HYG agar); culture volume was 500 µL per agar plate. See PK/PD experiment schemes in Table S3 (first study) and Table S6 (second study).
Selective agar was used because this agar had been prepared for plating the mouse lung homogenates associated with each infection time point. For CFU counts, + indicates too many colonies to accurately count; ~ indicates exact CFU count could not be determined due to merged/touching colonies; ---indicates not determined; and Contam. indicates bacterial of fungal contamination on agar that precluded CFU determination. CFU/mL (x) was log-transformed as log10 (x+1). LLOD, lower limit of detection. CFU counts for the following 10-fold dilutions: (shaded cell used to calculate CFU/mL) CFU/mL log 10 CFU/mL LLOD (log 10 CFU/mL) Table S9. Summary of M. bovis rBCG30 lung CFU counts in first PK/PD study. "Week 3 (all)" represents the combination of data from the Week 3 and Week 3 + 1.5 days time points. Data represent the mean (standard deviation) for each regimen at each time point. Unless otherwise noted, n = 5 mice per group per time point. See experiment scheme in

M. bovis rBCG30 mean (SD) log 10 CFU/lung at the following time points:
During treatment