Nathalie Gobeau (1), Mohammed Cherkaoui (1), Nicole Andenmatten (1), Didier Leroy (1), Nada Abla (1), Jörg Möhrle (1), Anne Kümmel (2) Francisco Javier Gamo Benito, Maria Jose Lafuente-Monasterio (3)
(1) MMV, Geneva, Switzerland, (2) IntiQuan GmbH, Basel, Switzerland (3) GSK, Tres Cantos, Spain
Introduction: Malaria is treated with combination therapies in order to minimize resistance. However, resistance is unavoidable and therefore the hunt for new treatments must continue. The objective of MMV is to develop new combinations that achieve 95% cure rates 28 days after a single-dose administration. Currently, the 28-day cure rates (ACPR28) of individual new molecules do not exceed 80%. Hence combinations must be identified so that 2 compounds together lead to a higher cure rate than each one individually. There are currently 8 antimalarial new molecules in the MMV portfolio which have already shown promising efficacy as monotherapy in humans – either in Controlled Human Malaria Infection (CHMI) studies or patient trials. From the large number of potential combinations of these candidates, the most promising ones need to be determined and progressed into late stage development.
Objectives: MMV aims to develop a strategy to screen a large number of combinations and select the most promising ones, i.e. the most likely to achieve the same efficacy with a single dose the as the current treatments which require three daily doses.
Methods: Individual compounds progress through drug development by going through a series of pharmacological models and by using PKPD models to identify the PKPD relationship and refine it at each stage. The first model is in SCID mice which are engrafted with human red blood cells and infected with P. falciparum, the most life-threatening malaria parasite specific to humans (1). The compound is administered at different doses when the parasitemia levels reach a predefined threshold. Blood concentrations and parasitemia are monitored for up to 2 months. A PKPD model is built in which the killing rate of the drug is related to its blood concentration by a sigmoidal Emax model – or a more complex model such as a turnover model, if needed. The second pharmacological model is a CHMI study where healthy volunteers are infected with P. falciparum and the drug is administered when parasitemia reaches a level which will not cause symptoms (2). Blood concentrations and parasitemia levels are measured for up to one month. Doses tested are subtherapeutic in order to determine the full PKPD relationship and in particular the minimum inhibitory concentration (MIC), the concentration below which parasites recrudesce. Finally, the compounds are tested in small cohorts of patients at different doses, high enough to not put the patients at risk.
Results: To screen combinations, a PBPK-PD model is generated to predict in humans the time-course parasitemia and calculate the clinical endpoint ACPR28, by using:
- The knowledge of the behavior of each compound in healthy volunteers or patients as monotherapy for both the PK and PD.
- The estimation of a possible interaction between the compounds in humans based on in vitro metabolism data and PBPK modelling: the PBPK model for each compound is validated with the human data – in patients if available or in healthy volunteers; the effect of the interaction on the PK profile of each combination partner is predicted by PBPK based on their victim and perpetrator properties.
- The estimation of the pharmacological interaction in SCID mice: several dosing regimens of the compounds as combination and monotherapy are tested. The general pharmacological interaction model (3) is applied to combine the existing monotherapy models and include a possible shift in Emax and EC50 due to the interaction between the drugs. This approach is general and can be adapted to more complex models, such as turnover models if needed. The shift estimated with the SCID dataset is then assumed to be the same in humans. The values of Emax and EC50 are, however, not translated from animals: the latest information obtained for the compounds as monotherapy – in challenge volunteers or patients – is used.
The example of the combination of MMV048 and DSM265 will be used for illustration purposes.
Conclusion: 12 combinations will be tested this year. It is expected that the current strategy will be able to identify at least the combinations which will lead to an ACPR28 higher than the ACPR28 of each individual agent, and at best allow to rank the different combinations according to their predicted ACPR28. This would then help prioritize the combinations that would progress into the CHMI study to confirm the predictions before going into patients.
References:
[1] Angulo-Barturen I et al., (2008) PLOS ONE 3(5): e2252. https://doi.org/10.1371/journal.pone.0002252
[2] McCarthy JS et al., Journal of Antimicrobial Chemotherapy (2016) 71(9):2620-2627. https://doi:10.1093/jac/dkw174
[3] Wicha et al, Nature Communications (2018) 8, 2129.
Reference: PAGE 27 (2018) Abstr 8739 [www.page-meeting.org/?abstract=8739]
Poster: Methodology - Other topics