Federico Romano (1) Jinghan Yu (1) Morris Muliaditan (2) Oscar Della Pasqua (1,3)
Institution (1) Department of Clinical Pharmacology and Therapeutics, University College London (London), UK, (2) Leiden Experts on Advanced Pharmacokinetics and Pharmacodynamics (LAP&P), Leiden, The Netherlands (3) Clinical Pharmacology Modelling & Simulation, GlaxoSmithKline, Brentford, UK
INTRODUCTION
A limitation in antitubercular drug development remains the empiricism based on which companion drugs and their dose rationale are selected for clinical testing. Even the most recent compounds approved for the treatment of tuberculosis seem to have a poor dose rationale. Model-based approaches can offer a method of discriminating whether the dose and exposure to an anti-tubercular compound is appropriate to ensure optimal antibacterial activity (i.e., bactericidal and sterilizing effects). Recently, the use of a parametric bacterial growth dynamics model enabled the characterisation of the contribution of companion drugs to the overall antibacterial activity of bedaquiline (B) in a mouse model of tuberculosis (TB) [1]. This semi-mechanistic model was developed to include fast-(F) and slow-(S) growing bacterial subpopulations, which represent the log-growth and stable phase of Mycobacterium. tuberculosis infection [1]. Here we use the approach to translate treatment effect and predict the early bactericidal activity (EBA) of B-containing regimens in humans.
METHODS
Published parameters of a bacterial growth dynamics model describing the antibacterial activity of B as monotherapy and in combination with pretomanid (Pa) and pyrazinamide (Z) in mice [1] were first scaled to humans using previously reported scaling strategies [2]. A population pharmacokinetic (PK) model developed by McCleay et al [5] was used to describe the steady state exposure of B in TB patients (n=111) using data from three clinical studies, TMC207-CL001 [2], NC-001 [3] and NC-003 [4]. In NC-001, B was tested using 400mg/day and in NC-003 200mg/day. For Pa-containing regimens, the PK of Pa was modelled in humans using data from two studies (CL-010 and CL-007) [6,7]. The PK of Z has been described by Muliaditan et al. [5], but data were not incorporated into the analysis as it was not used as backbone treatment. Acute and chronic assays conducted in mice [9-11] were used to estimate the antibacterial effect for Pa separately, and similarly scaled to humans taking into account changes in plasma protein binding. Predicted steady-state concentration in patients were then used in conjunction with the scaled system-specific parameters to predict the EBA over 14 days in drug-sensitive TB patients following treatment with B, B-Pa-Z, B-Pa, and B-Z. Model-based simulations were compared to the observed data to establish the predictive performance of the approach [2-4].
RESULTS
For Pa, the population clearance and central volume of distribution estimates were 3.9L/h and 107L, respectively. A delayed-effect compartment was required to describe the delayed antibacterial activity following monotherapy. The equilibrium rates KE0and KE1 were estimated to be 0.0018 and 0.007 1/h, respectively. Simulation scenarios including combination regimens showed greater median antimicrobial activity of B-Pa-Z [0.172 log10 CFU/ml/day] versus B-Z [0.128 log10 CFU/ml/day] versus B-Pa [0.102 log10 CFU/ml/day] versus 400mg/day B [0.098-0.104 log10 CFU/ml/day] over 0-14 days of treatment.
CONCLUSIONS
Despite the evolving knowledge about the use of model-based approaches for the evaluation of PK drug-drug interactions, limited attention has been given to the characterisation of pharmacodynamic interactions when developing drug combinations. The use of a bacterial growth dynamics model provides a robust basis for quantifying and predicting the overall antibacterial activity of combination therapy with multiple drugs. Moreover, this study shows that PKPD estimates in a murine model of infection can be used to predict the 14-day EBA of bedaquiline-containing novel combination regimens in humans.
References
[1] Muliaditan M, Della Pasqua O, Br J Clin Pharmacol, 2021. 87(1):140-151
[2] Diacon, AH, et al., Antimicrob Agents Chemother, 2013. 57(5): 2199-2203
[3] Diacon, AH, et al., Antimicrob Agents Chemother, 2020. 64(4): e02012 19
[4] Diacon, AH, et al., Am J Respir Crit Care Med, 2015. 191(8): 943-53
[5] McLeay SC, Vis P, van Heeswijk RP, Green B, Antimicrob Agents Chemother, 2014. 58(9): 5315-5324
[6] Diacon AH et al., Antimicrob Agents Chemother, 2010. 54: 3402–3407
[7] Diacon AH, et al., Antimicrob Agents Chemother, 2012. 56: 3027–3031
[8] Muliaditan M, Della Pasqua O.J Antimicrob Chemother. 2019 Nov 1;74(11):3274-3280
[9] Tyagi S, et al., Antimicrob Agents Chemother, 2005. 49(6): 2289-2293
[10] Tasneen R, et al., Antimicrob Agents Chemother, 2011. 55(12): 5485-5492
[11] Lenaerts AJ, et al., Antimicrob Agents Chemother. 2005. 49(6): 2294-301.
Reference: PAGE 29 (2021) Abstr 9859 [www.page-meeting.org/?abstract=9859]
Poster: Drug/Disease Modelling - Infection