Dynamic Mathematical Modeling of IL13-Induced Signaling in Hodgkin and Primary Mediastinal B-Cell Lymphoma Allows Prediction of Therapeutic Targets

Division of Systems Biology of Signal Transduction, DKFZ-ZMBH Alliance and Molecular Genetics, German Cancer Research Center, Heidelberg, Germany.
Cancer Research (Impact Factor: 9.33). 02/2011; 71(3):693-704. DOI: 10.1158/0008-5472.CAN-10-2987
Source: PubMed

ABSTRACT Primary mediastinal B-cell lymphoma (PMBL) and classical Hodgkin lymphoma (cHL) share a frequent constitutive activation of JAK (Janus kinase)/STAT signaling pathway. Because of complex, nonlinear relations within the pathway, key dynamic properties remained to be identified to predict possible strategies for intervention. We report the development of dynamic pathway models based on quantitative data collected on signaling components of JAK/STAT pathway in two lymphoma-derived cell lines, MedB-1 and L1236, representative of PMBL and cHL, respectively. We show that the amounts of STAT5 and STAT6 are higher whereas those of SHP1 are lower in the two lymphoma cell lines than in normal B cells. Distinctively, L1236 cells harbor more JAK2 and less SHP1 molecules per cell than MedB-1 or control cells. In both lymphoma cell lines, we observe interleukin-13 (IL13)-induced activation of IL4 receptor α, JAK2, and STAT5, but not of STAT6. Genome-wide, 11 early and 16 sustained genes are upregulated by IL13 in both lymphoma cell lines. Specifically, the known STAT-inducible negative regulators CISH and SOCS3 are upregulated within 2 hours in MedB-1 but not in L1236 cells. On the basis of this detailed quantitative information, we established two mathematical models, MedB-1 and L1236 model, able to describe the respective experimental data. Most of the model parameters are identifiable and therefore the models are predictive. Sensitivity analysis of the model identifies six possible therapeutic targets able to reduce gene expression levels in L1236 cells and three in MedB-1. We experimentally confirm reduction in target gene expression in response to inhibition of STAT5 phosphorylation, thereby validating one of the predicted targets.

1 Follower
23 Reads
  • Source
    • "e.g. [4] [5] [6] [7] [8] [9]. They are characterized by a functional relationship between the current state of the system x(t) ∈ R N at time point t and its time derivative ˙ x(t): ˙ x(t) = f (t, x(t), u(t), ζ) . "
    [Show abstract] [Hide abstract]
    ABSTRACT: In this work we present results of a detailed Bayesian parameter estimation for an analysis of ordinary differential equation models. These depend on many unknown parameters that have to be inferred from experimental data. The statistical inference in a high-dimensional parameter space is however conceptually and computationally challenging. To ensure rigorous assessment of model and prediction uncertainties we take advantage of both a profile posterior approach and Markov chain Monte Carlo sampling. We analyzed a dynamical model of the JAK2/STAT5 signal transduction pathway that contains more than one hundred parameters. Using the profile posterior we found that the corresponding posterior distribution is bimodal. To guarantee efficient mixing in the presence of multimodal posterior distributions we applied a multi-chain sampling approach. The Bayesian parameter estimation enables the assessment of prediction uncertainties and the design of additional experiments that enhance the explanatory power of the model. This study represents a proof of principle that detailed statistical analysis for quantitative dynamical modeling used in systems biology is feasible also in high-dimensional parameter spaces.
    Mathematical biosciences 04/2013; 246(2). DOI:10.1016/j.mbs.2013.04.002 · 1.30 Impact Factor
  • Source
    • "The Bayesian framework for the compartmental model analysis developed in the present work is directly applicable to a personalized dose assessment and the uncertainty quantification if a person-specific monitoring is requested. More generally, the presented methodology is suitable for any ODE-based model selection task, such as the modeling of protein signaling, gene regulation, or drug processing [47], nowadays frequently put forward in systems biology [48,49] or pharmacogenetics [50]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Background In radiation protection, biokinetic models for zirconium processing are of crucial importance in dose estimation and further risk analysis for humans exposed to this radioactive substance. They provide limiting values of detrimental effects and build the basis for applications in internal dosimetry, the prediction for radioactive zirconium retention in various organs as well as retrospective dosimetry. Multi-compartmental models are the tool of choice for simulating the processing of zirconium. Although easily interpretable, determining the exact compartment structure and interaction mechanisms is generally daunting. In the context of observing the dynamics of multiple compartments, Bayesian methods provide efficient tools for model inference and selection. Results We are the first to apply a Markov chain Monte Carlo approach to compute Bayes factors for the evaluation of two competing models for zirconium processing in the human body after ingestion. Based on in vivo measurements of human plasma and urine levels we were able to show that a recently published model is superior to the standard model of the International Commission on Radiological Protection. The Bayes factors were estimated by means of the numerically stable thermodynamic integration in combination with a recently developed copula-based Metropolis-Hastings sampler. Conclusions In contrast to the standard model the novel model predicts lower accretion of zirconium in bones. This results in lower levels of noxious doses for exposed individuals. Moreover, the Bayesian approach allows for retrospective dose assessment, including credible intervals for the initially ingested zirconium, in a significantly more reliable fashion than previously possible. All methods presented here are readily applicable to many modeling tasks in systems biology.
    BMC Systems Biology 08/2012; 6(1):95. DOI:10.1186/1752-0509-6-95 · 2.44 Impact Factor
  • Source
    • "CREB bZIP 40,000 rat PC12 pheochromocytoma cell line Conkright et al., 2003 E2A bHLH 30,000 murine primary pro B cells Zhuang et al., 2004 GR Receptor zinc finger 13,000 murine primary macrophage Salkowski and Vogel, 1992 Ikaros C2H2 zinc finger 250,000 murine thymocytes Bradley et al., 2003 NF-kB p65 Rel 250,000 murine macrophage Bradley et al., 2003 TR Receptor zinc finger 4,000 rat liver tissue Oppenheimer et al., 1974 Human AP-2a AP-2 200,000 HepG2 hepatocarcinoma cell line Egener et al., 2005 ER Receptor zinc finger 11,000; 260,000 ZR-75-1 breast cancer cell line; MCF-7 breast carcinoma cell line Reese and Katzenellenbogen, 1992 fos bZIP 4,000 NA De Togni et al., 1988 GR Receptor zinc finger 100,000 HeLa carcinoma cell line van Steensel et al., 1995 myc bHLH-ZIP 60,000; 140,000 MCF-7 breast carcinoma cell line; HCT-116 colon carcinoma cell line Nieddu et al., 2005 MyoD bHLH 50,000 RD rhabdomyosarcoma cell line S. Tapscott, personal communication NF-kB p65 Rel 120,000 TNFa stimulated T-leukemia cell line Hottiger et al., 1998 p53 p53 21,000; 160,000 LB19 B-lymphoblastoid cell line; MCF-7 breast carcinoma cell line Ma et al., 2005 PR Receptor zinc finger 200,000 T4D7 breast carcinoma cell line Nordeen et al., 1989 P-Smad2 SMAD 20,000 TGF-b stimulated keratinocyte cell line Zi et al., 2011 STAT2 STAT 150,000 INFa stimulated T-leukemia cell line Hottiger et al., 1998 STAT6 STAT 10,000; 200,000 CD19+ primary B cells; B cell lymphoma cell line Raia et al., 2011 Tcf-1 HMG 3,900,000 Jurkat T-leukemia cell line Van de Wetering et al., 1996 Developmental Cell 21, October 18, 2011 ª2011 Elsevier Inc. 617 Developmental Cell Perspective accessible DNA by electrostatic, sequence-independent interactions (K D $10 À6 M) (Lin and Riggs, 1975; von Hippel, 2004; von Hippel et al., 1974) and other molecules by sequencespecific interactions with many of tens of thousands of moderate and high-affinity recognition sites (K D < 10 À8 M) (e.g., Li et al., 2011). Methods such as FRAP, in vivo single-molecule analysis, and in vivo footprinting have been used to establish that indeed at least 90% of the molecules of each sequence specific DNA binding protein examined are bound to DNA in cells at any instant (Elf et al., 2007; Kao-Huang et al., 1977; Phair et al., 2004; Yang and Nash, 1995). "
    [Show abstract] [Hide abstract]
    ABSTRACT: To understand how transcription factors function, it is essential to determine the range of genes that they each bind and regulate in vivo. Here I review evidence that most animal transcription factors each bind to a majority of genes over a quantitative series of DNA occupancy levels. These continua span functional, quasifunctional, and nonfunctional DNA binding events. Factor regulatory specificities are distinguished by quantitative differences in DNA occupancy patterns. I contrast these results with models for transcription networks that define discrete sets of direct target and nontarget genes and consequently do not fully capture the complexity observed in vivo.
    Developmental Cell 10/2011; 21(4):611-26. DOI:10.1016/j.devcel.2011.09.008 · 9.71 Impact Factor
Show more