EUROPEANJOURNALOFCANCER46 (2010) 21– 32 available at www.sciencedirect.com journal homepage: www.ejconline.com Position Paper Modelling the genesis and treatment of cancer: The potential 5 role of physiologically based pharmacodynamics Jean-Louis Steimer a, Svein G. Dahl b, Dinesh P. De Alwis c, Ursula Gundert-Remy d, Mats O. Karlsson e, Jirina Martinkova f, Leon Aarons g,*, Hans-Ju¨ rgen Ahr h, Jean Clairambault i, Gilles Freyer j, Lena E. Friberg e, Steven E. Kern k, Annette Kopp-Schneider l, Wolf-Dieter Ludwig m, Giuseppe De Nicolao n, Maurizio Rocchetti o,In˜ aki F. Troconiz p a Novartis Pharma AG, Basel, Switzerland b Department of Pharmacology, Institute of Medical Biology, University of Tromsø, Norway c Eli Lilly & Company, Surrey, UK d Bundesinstitut fu¨ r Risikobewertung, Berlin, Germany e Uppsala University, Uppsala, Sweden f Charles University of Prague, Faculty of Medicine in Hradec Kralove, Hradec Kralove, Czech Republic g University of Manchester, Manchester, United Kingdom h Bayer Health Care, Leverkusen, Germany i INRIA, Paris, France j Medical Oncology Unit, Centre Hospitalier Lyon-Sud, Pierre-Be´nite, France k College of Pharmacy, University of Utah, USA l Department of Biostatistics, German Cancer Research Center, Heidelberg, Germany m Robert-Ro¨ssle-Klinik Oncology and Tumorimmunology, Berlin-Buch, Germany n Department of Computer Engineering and Systems Science, University of Pavia, Pavia, Italy o Accelera, Milan, Italy p Departamento de Farmacia y Tecnologı´aFarmace´utica, University Navarra, Pamplona, Spain ARTICLE INFO ABSTRACT Article history: Physiologically based modelling of pharmacodynamics/toxicodynamics requires an a priori Received 16 April 2009 knowledge on the underlying mechanisms causing toxicity or causing the disease. In the Received in revised form 30 context of cancer, the objective of the expert meeting was to discuss the molecular under- September 2009 standing of the disease, modelling approaches used so far to describe the process, preclin- Accepted 9 October 2009 ical models of cancer treatment and to evaluate modelling approaches developed based on Available online 1 December 2009 improved knowledge. 5 This paper is based in part on discussions at a COST B25 expert meeting held in Prague, Czech Republic, on 20–21st September 2007. Participating experts were Leon Aarons (UK), Hans-Ju¨ rgen Ahr (Germany), Jean Clairambault (France), Svein G. Dahl (Norway), Dinesh De Alwis (UK), Gilles Freyer (France), Lena Friberg (Sweden), Ursula Gundert Remy (Germany), Steve Kern (US), Anette Kopp-Schneider (Germany), Wolf-Dieter Ludwig (Germany), Jirina Martinkova (Czech Republic), Maurizio Rocchetti (Italy), Giuseppe De Nicolao (Italy), Jean-Louis Steimer (Switzerland), In˜ aki F. Troconiz (Spain). * Corresponding author: Address: University of Manchester, School of Pharmacy and Pharmaceutical Sciences, Oxford Road, M13 9PL Manchester, United Kingdom. Tel.: +44 161 275 2357; fax: +44 161 275 2396. E-mail address: [email protected] (L. Aarons). 0959-8049/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.ejca.2009.10.011 22 EUROPEANJOURNALOFCANCER46 (2010) 21– 32 Molecular events in cancerogenesis can be detected using ‘omics’ technology, a tool applied Keywords: in experimental carcinogenesis, but also for diagnostics and prognosis. The molecular Pharmacokinetics understanding forms the basis for new drugs, for example targeting protein kinases specif- Pharmacodynamics ically expressed in cancer. At present, empirical preclinical models of tumour growth are in Physiologically based modelling great use as the development of physiological models is cost and resource intensive. Cancerogenesis Although a major challenge in PKPD modelling in oncology patients is the complexity of Cancer therapeutics the system, based in part on preclinical models, successful models have been constructed Physiologically based models describing the mechanism of action and providing a tool to establish levels of biomarker Molecular mechanisms associated with efficacy and assisting in defining biologically effective dose range selection Intracellular pharmacokinetics– for first dose in man. To follow the concentration in the tumour compartment enables to pharmacodynamics link kinetics and dynamics. In order to obtain a reliable model of tumour growth dynamics Preclinical models and drug effects, specific aspects of the modelling of the concentration–effect relationship in cancer treatment that need to be accounted for include: the physiological/circadian rhythms of the cell cycle; the treatment with combinations and the need to optimally choose appropriate combinations of the multiple agents to study; and the schedule depen- dence of the response in the clinical situation. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction 2. Biological mechanisms of cancerogenesis Physiologically based pharmacokinetic models are designed 2.1. Molecular events in cancerogenesis to represent the human body and its physiology. The models allow the implementation of age- and disease-dependent In the traditional approach, the evaluation of the toxic prop- physiological changes enabling simulations and predictions erties of substances is based on phenotypic pathological of the kinetics under special circumstances which are exper- and histopathological characterisation of responses after imentally hardly accessible, e.g. in internal organs such as the exposure which does not allow insight into the underlying liver, and in special ages such as preterm babies. Up to now, toxic mechanism. In recent years, toxicogenomic techniques there are not many examples of a similar approach used to have been used to support the interpretation of the pheno- describe the dynamic part of the relationship between dose typic results by mode of action considerations. Furthermore, and observed effects. Models for pharmacodynamic/toxicody- gene expression techniques using microarrays have been ap- namic effect require a priori knowledge of the mechanistic ba- plied in toxicological studies with the aim of classifying sis of the disease process and at which step in the chain of chemicals using biomarkers based on gene expression 1 events the drug acts. The objective of the expert meeting changes. If the mode of action is understood and can be cap- was, in the context of cancer, to discuss the developments tured in a set of biomarkers, the detection of molecular events in: the molecular understanding of the disease process’ in would allow the prediction of selected toxic effects. Currently, the modelling approaches used so far to describe the disease the carcinogenic potential of chemicals is evaluated with ro- processes; preclinical models of cancer treatment; and thera- dent life time bioassays, which are time consuming and peutic options which have been recently developed based on expensive with respect to cost, number of animals and the improved understanding of the underlying processes. amount of compound required. Since the results of these 2- Thus, the meeting brought together experts involved in eval- year bioassays are not known until late during development uating and modelling steps in carcinogenesis, in the pharma- of new chemical entities, including drugs, and since the short ceutical development of new oncology agents or in the time test battery to assess genotoxicity, a characteristic of therapy of cancer. The discussions among the participants genotoxic carcinogens, is hampered by low specificity, the provided new insight into the contribution of physiologically identification of early biomarkers would be a big step forward. based modelling of pharmacodynamics to current and future Gene expression profiling on Affymetrix arrays has been cancer treatment. used to investigate the molecular events leading to a carcino- 2,3 The manuscript is organised as follows. In the first two genic response in rats. In a first study male rats were dosed sections the molecular basis of cancerogenesis, diagnosis with the genotoxic hepatocarcinogen N-nitrosomorpholine and treatment intervention together with their respective for 7 weeks followed by a treatment-free observation period 4 modelling approaches are presented. Section 4 describes of up to 50 weeks. Already shortly after start of the treat- how modelling and simulation is currently contributing to ment, significant alterations of various genes were observed, many stages and aspects in anticancer drug development: which belong to reasonable mechanistic pathways. Extension preclinical tumour growth models; management of drug (hae- of the observation period did not add much further informa- 5 mo-) toxicity; intra-tumour pharmacokinetics; physiological tion. Therefore, the study duration was limited to 14 d for aspects of tumour cells and predicting outcomes from drug further mechanistic studies and for the selection of early bio- combinations. markers. In a series of short-term in vivo studies, it was inves- EUROPEANJOURNALOFCANCER46 (2010) 21– 32 23 tigated whether carcinogens at doses known to induce liver set of independent validation compound profiles with up to tumours in the 2-year rat bioassay alter the expression of 88% accuracy. This may be considered as a proof of the con- characteristic sets of genes and whether these genes repre- cept that a classification of carcinogens based on short-term sent defined biological pathways. Male rats were dosed with studies is feasible.7 The usefulness of this approach has also five genotoxic and five non-genotoxic hepatocarcinogens.