Federico Romano, Salvotore D’Agate, Elisa Tacconi, Maria.J.Lafuente-Monasteria, Sven C. van Dijkman, Oscar Della Pasqua
University College London, Clinical Pharmacology and Therapeutics; GlaxoSmithKline (GSK), Diseases of the Developing World (DDW)
Objectives:
A semi-mechanistic pharmacokinetic-pharmacodynamic (PKPD) model was developed to help screen antimalarial compounds in the pre-clinical phase [1]. Assuming that the sensitivity of Plasmodium falciparum to mefloquine is comparable in both mice and humans, we attempted to scale drug- and disease-specific parameters from mice to humans as the basis for the prediction of parasitaemia in patients. The primary objective was to evaluate whether this model could make accurate quantitative parasite count predictions in humans based on humanised murine PKPD data. The predictive performance was assessed over the acute phase of infection (72 hours) and the rate of recrudescence over 28 days. Mefloquine was used as the investigative compound.
Methods:
Pre-clinical data were obtained from mice engrafted with human erythrocytes infected by Plasmodium. falciparum. The experimental protocol was ethically reviewed and carried out in accordance with the European Directive 2010/63/EEC and the GSK Policy on the Care, Welfare and Treatment of Animals. To develop a protozoa growth model, a semi-mechanistic drug-disease model was developed where the erythrocytic cycle was parameterised using six-compartments. The model allowed for natural regrowth of the parasite through feedback mechanisms using exponentials which can only decrease or increase and never reach zero. Thus, when treatment is stopped, the number of parasite cells described by the differential equation will naturally increase (signifying regrowth). A PK model was developed which adequately described mefloquine concentrations over the five dose levels which were tested. The PD model parameters were fixed to the estimated values obtained from the PD analysis. Drug effect was measured by potency (EC50) and EMAX, and parameterised for the clearance of human erythrocytes (Kout) and on the degradation rate of merozoites (Kdeath). For translational predictions, a human PK model published by [2] was substituted in, and the PD estimates from the final mice model were used to predict human parasite counts. A clinical study published by [4] was replicated using the individual body weights and baseline parasitaemia as covariates. Model predictions were compared with the observed clinical study findings. The analysis was performed in NONMEM V7.3 [2], whereas data manipulation was performed using R V3.0.1 and R Studio user interface. Model performance was assessed based on graphical and statistical criteria (RSE), goodness of fit and visual predictive checks (VPC).
Results:
Estimates of drug potency and efficacy for the acute stage for KOUT, which describes the clearance of human erythrocytes, and KDEATH, which describes the degradation rate of merozoites were: EC50KOUT (0.0213 ng/mL), EMAXKOUT (33.4), EC50KDEATH (0.0074 ng/mL) and EMAXKDEATH (27.6). IC50 Kdeath 13.6 [ng/mL], EMAX KDeath 7.19; IC50 Kout [ ng/mL ] 2.53; EMAX KOut 49.5; γ [Kout] 0.363 and γ [Kdeath] 0.627. For the prediction of recrudescence over 28 days, IC50 Kdeath 13.6 [ng/mL], EMAX KDeath 7.19; IC50 Kout [ ng/mL ] 2.53; EMAX KOut 49.5; γ [Kout] 0.363 and γ [Kdeath] 0.627. Assuming a limit of detection for microscopy of 20 parasites per μl, the predicted parasitaemia/cure at 72 hours is similar to the observed data (47 vs. 63%) in the clinical study by Looareesuwan et al (1999) 2. Over 28 days, recrudescence was predicted in 7 cases vs. 11 observed cases in the original study.
Conclusions:
The use of a model-based semi-mechanistic approach to scale pharmacodynamics from a murine model seems to enable the prediction of treatment response to mefloquine in patients. Whilst these findings may not be immediately generalisable, this exercise provides the groundwork for the translation of drug and disease characteristics from a humanised murine model to malaria infection in patients. Possible future uses for the model would be a tool for personalised predictions of parasite counts.
References:
[1] Tacconi et al. Model-based evaluation of anti-parasital compounds for the treatment of Plasmodium Falciparum. Pharmacology conference. 2018
[2] NONMEM v.7.3 (ICON Development Solutions, Ellicott City, MD).
[3] Reuter et al. Population pharmacokinetics of orally administered mefloquine in healthy volunteers and patients with uncomplicated Plasmodium falciparum malaria. Journal of Antimicrobial Chemotherapy, Volume 70, Issue 3, 1 March 2015, Pages 868–876,
[4] Looareesuwan et al. Efficacy and safety of atovavauone/proguanil compared with mefloquine for treatment of acute plasmodium falciuparum malaria in Thailand. Trop. Med. Hyg., 60(4), 1999, pp. 526–532 Copyright q 1999 by The American Society of Tropical Medicine and Hygiene.
Reference: PAGE 28 (2019) Abstr 9142 [www.page-meeting.org/?abstract=9142]
Poster: Drug/Disease Modelling - Infection