Raia et al., (2010). Dynamic Mathematical Modeling of IL13Induced Signaling in Hodgkin and Primary Mediastinal BCell Lymphoma Allows Prediction of Therapeutic Targets.
April 2016, model of the month by Thawfeek Mohamed Varusai
Introduction: Lymphomas are tumours of the lymphatic system and are quite heterogeneous in nature. There are several types of clinical malignancies of lymphomas, which include primary mediastinal Bcell lymphoma (PMBL) and classical Hodgkin lymphoma (cHL). The Janus kinase signalling pathway (JAK/STAT) is known to signal a range of cytokines and growth factors, which primarily trigger the downstream transcriptional process, thereby regulating several vital cellular functions such as cell proliferation and apoptosis. Consequently, the JAK/STAT pathway is found to be aberrant in many cancers including blood malignancies. In this paper, the authors use a combination of experimental and mathematical modelling approach to study the role of JAK/STAT signalling pathway in PMBL and cHL classes of lymphomas [1]. Figure 1The process diagram of MedB1 dynamic signaling network model consisting of reactions (arrows) with enzymatic, mass action, or custom kinetics. Roundheaded arrows indicate reaction catalysis, whereas barended arrows reaction inhibition. IL13 is used as input function of the system. Reactions and species colored in gray are omitted in the L1236 model. Figure taken from [1]. Motivation: The JAK/STAT signalling pathway is known to be deregulated in PMBL and cHL cases by being constitutively active. Pathway activating factors in PMBL and cHL include hyperphosphorylated STAT5 and STAT6 proteins, overexpression of JAK2 gene and stimulation of the JAK/STAT pathway by the secreted interleukin 13 (IL13). Some of the pathway aberrations in lymphoma are less clear such as the role of nonmutated negative regulators like SOCS3, constitutive expression of SHP1 and the differential roles of STAT5 and STAT6. For these reasons, the authors consider JAK/STAT signalling as a promising pathway to study in the context of lymphomas. Since the system involved is complex and nonlinear, a quantitative dynamic modelbased approach is required. Previous JAK/STAT pathway modelling attempts are mainly literaturebased or focused on nonpathological systems. This reduces the predictive power of the resultant model for pathological conditions. Very few other studies include empirical and modelling approach to understand the JAK/STAT pathway. The authors, therefore, use a combination of experiments and mathematical modelling in PMBL and cHL derived cell lines for their study. Mathematical Modelling: The authors describe the JAK/STAT pathway for PMBL and cHL conditions using a mathematical model. MedB1 cells (derived from PMBL) and L1236 cells (derived from cHL) were used to generate the experimental data required to describe the corresponding models. The two models were generated slightly different from each other to accommodate the different experimental observations in the two cell types. The MedB1 model has few additional components than L1236 model  the IL13Ra2 decoy receptor (DecoyR) and the SOCS3mediated negative feedback (Fig.1). Ordinary differential equations were used to represent the reactions. The authors use data from the time course experiments from IL13 treated cell lines to estimate the model parameters. Sensitivity analysis has been performed on the model parameters to identify the parameters that can serve as therapeutic interventions. Subsequently, the authors have experimentally validated a potential drug target in the system. Results: The heterogeneous nature of lymphoma have been emprically established by identifying the molecular stoichiometric differences in the JAK/STAT signalling components in MedB1 and L1236 cell lines. Experimental data show that the signalling pathway is more aberrant in L1236 cells than in MedB1, suggesting that the former is more aggressive than the latter. Experiments also shed light on the differential role of STAT family members in IL13 signalling – STAT5 is more responsive than STAT6 in both MedB1 and L1236 cell types. Databased mathematical modelling of the JAK/STAT pathway was used to gain a quantitative understanding of two systems. Sensitivity analysis of the models reveal that L1236 cells are more susceptible to perturbations than MedB1. Furthermore, sensitivity analysis show potential intervening points in the pathway that can be exploited for drug targeting in the lymphoma. Finally, the authors empirically validate model predictions by testing a potential drug target that is common to both lymphomas – STAT5 phosphorylation. Predictive Power of the Model: The predictive power of a mathematical model depends upon how well the model represents the real biological system of interest. Parameter values are key deciding factors of the model predictive power. In this study, the authors take several measures to accurately estimate the model parameters in the MedB1 and L1236 systems.
Therefore, the models can be claimed to have a high predictive power for the MedB1 and L1236 systems. This is corroborated by the observation that drug inhibition experiments on the JAK/STAT pathway in cells match with the sensitivity analysis of the models. However, to truly gain insight on the differential impact of tumour cells relative to normal cells, a JAK/STAT pathway model of a normal B cell would be required. The authors admit this but were not able to perform the task due to technical difficulties. Significance of the Model: Using mathematical models to understand the JAK/STAT pathway has been a fruitful approach in this study. As mentioned in the discussion section of the paper, the models reveal important biological knowledge about the JAK/STAT network. Model simulations show that IL13Ra2 decoy receptor has more impact on IL13 signalling in MedB1 cells than L1236 cells despite the fact that the expression of the receptor is higher in the latter than the former. Another observation in the study is that the empirical STAT5 phosphorylation peak could not be fully represented by the L1236 model, suggesting the involvement of nonmodeled factors such as cross talks from other pathways. Obtaining such information about the biological system using only experiments may either not be possible due to technological limitations or may involve laborious work. Scientific Value Added: The impact of a study on the scientific community is proportional to the degree of usage of its results by other researchers. The value that this study adds is as follows:
Bibliographic references
