Leire Ruiz Cerdá (1,2), Eduardo AsÃn Prieto (1,2), Annabel Blasi (3), Joan Carles Reverter (4) and Iñaki Trocóniz (1,2)
(1) Pharmacometrics & Systems Pharmacology, Department of Pharmacy and Pharmaceutical Technology, University of Navarra, Pamplona, Navarra, Spain. (2) Institute for Health Research (IdiSNA), Pamplona, Navarra, Spain. (3) Anesthesia Department. Hospital Clinic. Barcelona. Institut per a la recerca biomedica Agusti Pi i Sunyé (IDIBAPS). (4) Haemostasis and haemotherapy. Hospital Clinic. Barcelona.
Objectives: Several System Pharmacology (SP) models for the Coagulation Network are currently available in the literature (1-5). This type of models allows simulating the time profiles of the different components of the coagulation cascade and therefore, emulating the endpoints of several coagulation tests: (i) Thrombin Generation Assay (TGA), which captures Thrombin (FIIa) profiles, (ii) Prothrombin Time (PT) and (iii) activated Partial Thromboplastin Time (aPTT), which provide insight of the time needed to form the fibrin (FIIa) clot by the extrinsic or intrinsic pathway, respectively. In this regard, an unavoidable step is the evaluation of these SP models and their capacity to mimic the patho-physiological behavior by comparison with experimental data (6). Altogether, the objective of this work once the most relevant models were selected and implemented is to perform a validation exercise beyond the original experimental scenarios.
Methods: Two models were selected from the literature characterizing the entire coagulation network based on the inclusion of relevant components and reactions: Wajima, et al. (3) and Nayak, et al. (4). Firstly, these models were implemented using MATLAB’s (v. 2017a) SimBiology 5.6 toolbox, including model components, reaction rates and equations. Importantly, a special effort was made in order to define units and make the models comparable. After the implementation, the figures shown in each study were reproduced to check and validate the implementation. Secondly, experimental data from normal individuals and trauma patients were collected from Menezes, et al. (6). These data included: (i) percentage of activation of blood protein factors II, V, VII, VIII, IX, X, and ATIII in each normal and trauma sample, (ii) TGA, (iii) PT and (iv) aPTT. For all the patients TGA, PT and aPTT were simulated using as initial condition the blood factors percentage reported. PT and aPTT were calculated as the time at which the 30% of the fibrinogen (Fg) was transformed to FIIa (7). Initial individual conditions for factors and proteins that were not reported were assumed equal to those listed in each of the selected models. Time course of FIIa and PT and aPTT were simulated and compared.
Results: The two models selected from the literature were satisfactorily implemented, as shown by the exact reproduction of the results presented in both manuscripts. In addition simulations of PT and aPTT metrics obtained using levels of several factors provided in Menezes et al. (6) for healthy individuals agreed well with the experimental observations. When a similar exercise was performed for the case of patients suffering from trauma, the above mentioned descriptors showed a relative error over 30%. When the time course of different coagulation cascade factors were simulated, discrepancies were found between the profiles obtained from the different models. Once a sensitivity analysis was performed, it was clearly seen that longitudinal profiles were very sensitive to initial conditions driven the experimental conditions, which might explain the discrepancies encountered. Interestingly, the sensitivity of the area under the FIIa concentration-time curve (AUC) to the initial stages of the coagulation process is very low and the discrepancies seen in FIIa concentration profile were not reflected in the corresponding AUC, which is the measure driving PT and aPTT.
Conclusion: Both selected models described very well data gathered from healthy individuals. However, this was not the case at least for patients suffering from trauma. Differences either in other non-measured elements from the coagulation cascade, or in model parameters between healthy subjects and patients are likely to be the reasons responsible for poor model performance, highlighting the need to focus on disease related process alterations when developing systems pharmacology. For outcomes obtained from ex-vivo sample manipulation a very detailed description of the experimental setting is required to ensure model reproducibility.
References:
[1]. Hockin MF, Jones KC, Everse SJ, Mann KG. A model for the stoichiometric regulation of blood coagulation. J Biol Chem. 2002;277(21):18322–33.
[2]. Chatterjee MS, Denney WS, Jing H, Diamond SL (2010) Systems Biology of Coagulation Initiation: Kinetics of Thrombin Generation in Resting and Activated Human Blood. PLOS Computational Biology 6(9): e1000950.
[3]. Wajima T, Isbister GK, Duffull SB. A comprehensive model for the humoral coagulation network in humans. Clin Pharmacol Ther [Internet]. Nature Publishing Group; 2009;86(3):290–8.
[4]. Nayak S, Lee D, Patel-Hett S, Pittman DD, Martin SW, Heatherington AC, et al. Using a Systems Pharmacology Model of the Blood Coagulation Network to Predict the Effects of Various Therapies on Biomarkers. CPT Pharmacometrics Syst Pharmacol. 2015;4(7):396–405.
[5]. Lee D, Nayak S, Martin SW, Heatherington AC, Vicini P, Hua F. A quantitative systems pharmacology model of blood coagulation network describes in vivo biomarker changes in non-bleeding subjects. J Thromb Haemost. 2016;14(12):2430–45.
[6]. Menezes AA, Vilardi RF, Arkin AP, Cohen MJ. Targeted clinical control of trauma patient coagulation through a thrombin dynamics model. Sci Transl Med. 2017;9(371):1–12.
[7]. Kogan, A.E., Kardakov, D.V. & Khanin, M.A. Analysis of the activated partial thromboplastin time test using mathematical modeling. Thromb. Res. 101, 299–310 (2001).
Reference: PAGE 27 (2018) Abstr 8748 [www.page-meeting.org/?abstract=8748]
Poster: Drug/Disease Modelling - Other Topics