Aline Fuchs (1), Mohammed Cherkaoui Rbati (1), Anne Kuemmel (2), Jeremy Burrows (1), Didier Leroy (1), Azrin N. Abd-Rahman (4), Jörg J. Möhrle (1,3), Nathalie Gobeau (1)
(1) Medicines for Malaria Venture, Route de Pré-Bois 20, 1215 Meyrin, Geneva, Switzerland (2) IntiQuan GmbH, Spalenring 150, 4055 Basel, Switzerland (3) University of Basel, Basel, Switzerland (4) QIMR Berghofer Medical Research Institute, 300 Herston Rd, Brisbane, QLD 4006, Australia
Introduction The development of resistance poses a significant problem in the treatment of Plasmodium falciparum malaria, the deadliest of the five Plasmodia species infecting humans; and new medicines are urgently required. Medicines for Malaria Venture (MMV) and partners have set up in vitro assays, animal experiments and volunteer infection studies using P. falciparum, to help drug development and evaluation of candidate compounds. To aid the selection of the most promising compounds for clinical development, PKPD modelling is used to predict from preclinical efficacy data and available human PK information the human efficacious dose.
To validate the PKPD prediction method from preclinical to humans, a retrospective analysis of five compounds that have been tested in humans was performed. This evaluation focused on the greater challenge of pharmacodynamic prediction; human PK was not predicted from animal data.
Objective To assess the performance of human efficacious dose predictions from preclinical PKPD models.
Methods The efficacious dose was defined as that which killed 6 to 9 log(10) units of parasites, a minimum threshold that must be met for a clinical candidate at a well-tolerated dose [1].
For each compound the efficacious dose was calculated by constructing a PKPD model based on an in vitro Parasite Reduction Rate assay (PRR) [2] and P. falciparum infected highly immunodeficient mice (NSG) mouse experiment [3]. The PKPD model divided the parasite population into two compartments: one representing living, the other the dead parasites; the in vivo parasite assay potentially measured both. The differential equation describing the dynamics of living parasites included a net growth rate plus a sigmoid Emax function describing how the compound concentration stimulated parasite killing. The differential equation describing the dynamics of dead parasites had said kill rate as inflow plus a first-order elimination of clearing dead parasites from the body.
Unlike the in vivo assay, the in vitro PRR assay measured the number of viable parasites over time in presence of various compound concentrations. The Emax value for each compound was calculated as the slope of the viable parasite kill curve at the highest compound concentration, minus the in vitro net parasite growth rate. In the in vivo experiments mice were infected with P. falciparum and subsequently administered candidate compounds at different doses. Compound concentrations and parasitemia were measured over time. A two-stage PKPD model-based analysis was used: first a PK model was fitted then PD parameters estimated with fixed PK parameters. The potency (EC50, Hill coefficient) of the Emax sigmoidal model and the clearance of dead parasites were estimated from mouse data. The Emax value was fixed to that obtained from in vitro PRR data.
The PKPD model was used to make predictions for humans: the mouse PK model was replaced by a human version; the clearance parameter was set to a high value. The minimum dose to clear 6- and 9-log(10) parasites was calculated. These predictions were then compared with those obtained with the best selected PKPD model built directly from human volunteer infection studies.
Results The approach combining in vitro and mouse data gave more consistent and accurate dose predictions for humans compared with other approaches relying only on animal data [4]. The Emax parameters were within a 2-fold margin between in vitro and human. Doses predicted based on 6 log(10) parasite clearance were within a 2-fold margin between mouse and human. The same results were obtained for doses predicted based on 9 log(10) parasite reduction except for one compound for which there was a 4-fold deviation between predicted and observed.
Conclusion This work has improved the MMV methodology for human dose prediction of antimalarial compounds from preclinical data. To, potentially, reduce the number of animal experiments and increase efficiency (cost, time, accuracy and precision) ongoing work is investigating if further in vitro experimental data could also predict the EC50 and Hill coefficient. The next step will be to ensure that the PKPD model derived from healthy volunteer infection study data is representative of patient data.
References: [1] Burrows J et al. Malar J (2017). New developments in anti-malarial target candidate and product profiles
[2] Sanz LM et al. Plos One (2012). P. falciparum iIn vitro killing rates allow to discriminate between different antimalarial Mode-of-Action
[3] Angulo-Barturen I et al. Plos One (2008). A Murine Model of falciparum-Malaria by In Vivo Selection of Competent Strains in Non-Myelodepleted Mice Engrafted with Human Erythrocytes
[4] Fuchs A et al. PAGE 28 (2019) Abstr 8998 [www.page-meeting.org/?abstract=8998]. Assessment of translation of PKPD relationship from animal to human for malaria compounds
Reference: PAGE () Abstr 9340 [www.page-meeting.org/?abstract=9340]
Poster: Oral: Drug/Disease Modelling