Chiara Fornari (1), Lenka Oplustil O’Connor (2), Carmen Pin (1), James W.T. Yates (3), S. Y. Amy Cheung (4), Duncan I. Jodrell (5), Jerome T. Mettetal (6), and Teresa A. Collins (1)
(1) Safety and ADME Translational Sciences, Drug Safety and Metabolism, IMED Biotech Unit, AstraZeneca, Cambridge, UK (2) Oncology Translational sciences, IMED Biotech Unit, AstraZeneca, Cambridge UK (3) DMPK, Oncology, IMED Biotech Unit, AstraZeneca, Cambridge, UK (4) Quantitative Clinical Pharmacology, Early Clinical Development, IMED Biotech Unit, AstraZeneca, Cambridge, UK (5) CRUK Cambridge Institute, Li Ka Shing Centre, University of Cambridge, Cambridge, UK (6) Safety and ADME Translational Sciences, Drug Safety and Metabolism, IMED Biotech Unit, AstraZeneca, Boston, USA
Introduction: Balancing anti-tumour efficacy with safety is an ongoing challenge in oncology. Drug-induced myelosuppression is a common dose limiting toxicity of cancer treatments, for both cytotoxic drugs, targeted therapies, and combinations [1]. Mathematical modelling has proven to be a powerful aid in understanding the mechanisms of drug-induced haematotoxicity, scaling results from animal models to humans, and designing optimized treatment regimens [2]. However, major limitations with existing approaches are that (i) due to the invasiveness of bone marrow (BM) biopsy, myelosuppression is inferred from blood counts, especially in humans, extrapolating to the unseen effects in the BM and, (ii) often, each blood cell type (e.g. neutrophils, platelets) is modelled independently[3]. In order to close these gaps and understand the heterogeneity observed in clinical outcome, this routine data obtained from blood must be supplemented with additional information, such the one from in vivo haematopoiesis systems [4]. Then, systems models help to understand differences in time-course and to characterize the underlying biology, that may otherwise be overlooked and, ultimately, they can aid translation from the pre-clinical studies to human.
Objectives: To develop a novel mechanistic model of carboplatin-induced myelotoxicity that can capture the overall myelosuppression profile in rat, i.e. from progenitor cells in the BM to different types of cells in the blood (neutrophils, monocytes, platelets, reticulocytes, and red blood cells).
Methods: Carboplatin-induced myelotoxicity was derived using in vivo rat data following different exposures to carboplatin (30mg/Kg on day 1; 40mg/Kg every 14 days for multiple cycles). More precisely, we analysed time-courses of neutrophil, monocyte, platelet, reticulocyte and red blood cell counts together with BM cellularity profiles (multi potent progenitor, common myeloid progenitor, and megakaryocyte-erythrocyte progenitor counts). The model is a system of ordinary differential equations, implemented in Matlab, using the Simbiology platform. The same platform was also used to perform model fits to data and simulations.
Results: A new systems pharmacology model was developed to capture the essential features of haematopoiesis, such as the role of stem cells, the process of differentiation into multiple lineages, and the interaction among different feedback mechanisms, [5]. Moreover, we modelled haematopoiesis as a whole system, where all lineages come from a common set of progenitors. Then, we fit this model to data from in vivo haematopoiesis systems, capturing the dose and regimen dependent Carboplatin-induced myelosuppression profile in rats. The model described how variations in the BM are mechanistically linked to those in the blood, and how the interaction among inputs from different lineages modulate progenitor proliferation and differentiation, giving rise to the observed changes in BM cellularity and blood cell counts.
Conclusion: Drug-induced myelosuppression remains a significant issue in oncology. We reviewed existing mathematical approaches for studying haematopoiesis and bone marrow toxicity, identified gaps in current understanding, and made future recommendations to advance further this vital field of safety research. In particular, a major issue is obtaining observations from the BM, especially in humans, due to the invasiveness of BM biopsy. When this upstream information is not available, modellers generate hypotheses about the unseen drug effects, based only on drug concentration and blood counts. We developed a novel systems pharmacology approach based on in vivo rat data to link carboplatin-induced toxicity in the BM with effects in the blood, thus following the overall myelosuppression dynamics from progenitors to mature cells. The benefits of a model such as this include (i) improved physiological understanding of drugs effects, (ii) less inference of upstream outcomes from blood cell counts, and (iii) better characterization of the correlations existing among different types (and grades) of haematotoxicity. Lastly, this work represents a first step to establish animal model to build confidence in first time dose in man for translation. Therefore, by developing this further, we will learn more about the impact of therapeutics on the myeloid system and, potentially, develop a tool to translate pre-clinical results and guide treatment scheduling in the clinic.
References:
[1] Barreto et al., Antineoplastic agents and the associated myelosuppressive effects: a review, J Pharm Pract, 2014
[2] de Vries Schultink et al., Pharmacodynamic modelling of adverse effects of anti-cancer drug treatment. Eur J Clin Pharmacol, 2016
[3] Friberg LE and Karlsson MO, Mechanistic models for myelosuppression, Invest New Drugs, 2003
[4] Kakiuchi et al., Flow cytometric analyses on lineage-specific cell surface antigens of rat bone marrow to seek potential myelotoxic biomarkers: status after repeated dose of 5-fluorouracil, J Toxicol Sci, 2004
[5] Manesso et al., Dynamical modelling of haematopoiesis: an integrated view over the system in homeostasis and under perturbation. J R Soc Interface, 2013
Reference: PAGE 27 (2018) Abstr 8507 [www.page-meeting.org/?abstract=8507]
Poster: Drug/Disease Modelling - Other Topics