Chenyao Liu 1,4, Evangelos Karakitsios 1, Silvia Grandoni 1, Salvatore D’ Agate 1, Samira Boarbi 2, Mehi Kiass 2, Pierre-Yves Adnet 2, Vanessa Mathys 2, Robert Bates 3, Laura Guijarro-Lopez 3, Mara Martinez-Martinez 3, Carlos Rey 3, Modesto Remunian-Blanco 3, Angel Santos-Villarejo 3, Oscar Della Pasqua 1,4
1 National Research Council (CNR) (Rome, Italy), 2 Unit "Tuberculosis & Mycobacteria", Human Bacterial Diseases Service, Sciensano (Brussels, Belgium), 3 GSK Global Health Medicines (Tres Cantos, Madrid, Spain), 4 University College London (London, UK)
Objectives: Ethionamide (Eto) is used in combination regimens for pulmonary multidrug-resistant tuberculosis (MDR-TB), with known poor tolerability. Alpibectir is currently under development1 as a booster of Eto at lower doses, enhancing Eto efficacy with less side effects. The combination of alpibectir and Eto, often referred as boosted ethionamide (bEto), exhibits clinical potential for TB treatment. To support clinical research of bEto combinations, preclinical research was conducted in a stepwise manner: a satellite PK study in healthy BALB/c mice followed by a CFU reduction experiment in infected BALB/c mice . The ultimate objective of this work is to optimize the design of CFU reduction study in the mouse, in order to subsequently develop pharmacokinetic-pharmacodynamic (PKPD) models for the assessment of concentration-effect relationships of drug combinations aimed at the evaluation of overall antibacterial activity.
Methods: In the satellite PK experiment, combinations including bedaquiline (B), pretomanid (Pa), moxifloxacin (M), rifapentine (P), isoniazid (H), pyrazinamide (Z) and ganfeborole (GSK656) were tested: BPabEto, BPaMbEto, PHZM, BPa656bEto and PZMbEto. Healthy Balb/c mice (Ntotal = 75) were administered anti-TB combinations for five consecutive days and blood samples were collected for PK analysis on the final day, reflecting steady state. All animal studies were ethically reviewed and carried out in accordance with European Directive 2010/63/EU. Population pharmacokinetic (popPK) models for each compound were developed using pooled data across combinations in NONMEM. Drug-drug interactions (DDIs) were assessed as covariate effects on PK parameters. Post-hoc steady-state exposure metrics derived from the individual popPK models were compared with target human exposure. Subsequently, dosing of each compound was optimised to achieve consistent exposure across groups and match target human exposure. Furthermore, the sampling scheme for the CFU reduction experiments was established using a D-optimal design to capture individual PK profiles. In particular, the design prioritised precise estimation and inter-individual variability of clearance for each compound, ensuring accurate individual exposure estimates to support subsequent exposure–CFU reduction analyses. This work has received support from the Innovative Medicines Initiatives 2 Joint Undertaking (grant No 853989).
Results: popPK models were developed as one-compartmental models (Pa, P, Z, GSK656, alpibectir) or two-compartmental models (B, M, Eto) with first-order absorption and first-order clearance. No DDIs across combinations were identified for bEto. Considering human exposure and practical feasibility, following dose was proposed for CFU experiment: B 25mg/kg, Pa 40 mg/kg, M 100mg/kg, Z 150mg/kg, isoniazid 10 mg/kg, GSK656 8mg/kg, alpibectir 1.5 mg/kg, Eto 25mg/kg. Optimal sampling of CFU experiment was identified considering ethical and laboratory capacity while capturing PK profile. Ten mice per group were designed to be sacrificed at two time-points (weeks 5 and 8, Nper timepoint = 5) for CFU count, while blood concentrations were measured at 3 time-points after five consecutive doses (weeks 3, 4 and 6) to support development of the final PK-PD model. Due later administration of bEto at 2h, parallel sampling schemes were applied to better capture PK profiles of bEto. For groups without bEto, sampling timepoints were proposed: 0.25, 1.08, 2.0, 4.0, 6.0 and 24 hours post-dosing; for groups with bEto, sampling at 2.08, 2.25, 3.0, 5.0, 6.25 and 24 hours post-dosing were proposed. The resulting sampling schemes demonstrated good parameter precision, with relatively low relative standard errors.
Conclusion: A satellite PK study in healthy mice was performed to optimize CFU-reduction experiment of bEto-containing combinations in TB-infected mice, accounting for potential DDIs and alignment with target human exposures. Addition of bEto to the evaluated anti-TB combinations was unlikely to produce relevant DDIs. Based on the developed popPK models and D-optimality criteria, human-equivalent dosing and an optimal sampling design was established for subsequent CFU study, which is currently underway. The final PKPD models will support assessment of concentration-effect relationships aimed at evaluation of overall antibacterial activity of bEto-included combinations.
References:
1. Pieren M, Abáigar Gutiérrez-Solana A, Antonijoan Arbós RM, et al. First-in-human study of alpibectir (BVL-GSK098), a novel potent anti-TB drug. J Antimicrob Chemother. 2024;79(6):1353-1361. doi:10.1093/jac/dkae107
Reference: PAGE 34 (2026) Abstr 12273 [www.page-meeting.org/?abstract=12273]
Poster: Drug/Disease Modelling - Infection