(M-098) Population pharmacokinetic modeling of TBA-7371 in healthy participants describes apparent auto-induction of clearance and predicts PK of participants with Mycobacterium tuberculosis

Shayne Watson, MSc – Clinical Pharmacology and Pharmacometrics Scientist, Bill & Melinda Gates Medical Research Institute; Micha Levi, PhD – Clinical Pharmacology Leader, Bill & Melinda Gates Medical Research Institute; David Holtzman, MD – Clinical Development Leader, Bill & Melinda Gates Medical Research Institute; Aparna Anderson, PhD – Portfolio Statistics Leader, Bill & Melinda Gates Medical Research Institute; Charles Wells, MD – Head of Therapeutics Development, Bill & Melinda Gates Medical Research Institute

Objectives:
Develop a population pharmacokinetics (popPK) model that can characterize the apparently complex PK of anti-Mycobacterium tuberculosis (TB) compound, TBA-7371, in healthy participants and participants with active TB.

Methods:
Data from a single- and multiple- ascending dose (SAD and MAD) Phase 1 trial were used to fit a preliminary popPK model for PK of TBA-7371 in healthy participants. Visual assessment of the concentration time-course data revealed complexities in the PK that needed to be addressed in the course of modeling: Accumulation of trough concentrations in the MAD trial were observed for the first few days of dosing followed by a decrease and leveling-off of trough concentrations, an apparent auto-induction of clearance. In addition, PK from higher SAD doses exhibited possible non-linear (saturable) elimination.

A limited set of covariates were considered in the modeling:
• The effect of body weight on distribution and elimination
• The effects of fed vs. fasted state and dose on absorption rate and extent.

Model-building proceeded as per standard practice [1]. Standard goodness-of-fit diagnostics and visual predictive checks (VPC) were used to guide the model-building, with care towards addressing the noted PK complexities.

The model was used to explore potential dosing regimens for a Phase 2 trial in participants with active TB. Once Phase 2 data were available, predictions were compared with data as an external validation VPC. The model was then updated to include both Phase 1 and Phase 2 data.

Results:
PK in healthy participants was well characterized using a 2-compartment model with first-order absorption, parallel linear and non-linear elimination (via Michaelis-Menten kinetics) and a hypothetical enzyme compartment to drive auto-induction of the non-linear elimination. Levels of the hypothetical enzyme were modeled analogously to indirect-response modeling of any PD endpoint but were used to scale (induce) clearance.

The effect of body weight was modeled via standard allometric exponents on clearance and volume parameters. Absorption was slower for fed (vs fasted) administration and for higher doses. No effect on extent of absorption was observed.

PK for selected Phase 2 dosing regimens were simulated using the preliminary popPK model. After completion of the Phase 2 trial, the preliminary model was observed to have predicted the Phase 2 data well. Model parameters were then updated based on combined Phase 1 and 2 data.

Conclusions:
PK of TBA-7371 in healthy participants was well characterized using a 2-compartment model with first-order absorption, parallel linear and non-linear elimination and a hypothetical enzyme compartment to drive auto-induction of the non-linear elimination. The preliminary model also described the PK in participants with active TB. The model was deemed fit for the purpose of informing selection of dosing regimens for further study.

Citations: [1] Nguyen TH, Mouksassi MS, Holford N, Al-Huniti N, Freedman I, Hooker AC, et al. Model Evaluation Group of the International Society of Pharmacometrics (ISoP) Best Practice Committee. Model evaluation of continuous data pharmacometric models: metrics and graphics. CPT Pharmacometrics Syst Pharmacol. 2017 Feb;6(2):87-109