PMC

Flow chart summarizing the workflow described in the introduction.

Proposed regulation of cell death in leukemia. In CEM C7 cells, upon GC treatment, GR becomes activated and alters Bim and GR transcription, potentially through AP-1 and Erg recruitment respectively. This may be cell specific as such recruitments were not seen in CEM C1 cells. Other factors such as MAPK signaling (i.e., JNK or P38) and c-myb may play a role in regulating GR, AP-1, and Bim.

Topology of GR/Erg models in CEM C7–14 cells. Schematic representation of GR/Erg pathway. The nature of the topologies was based on previously established direct GR target model (Chen et al., ). Model 3 and Model 4 are similar and only differ by the regulation of GR, in Model 3 GR regulation is controlled by GR itself (A) whereas in Model 4 GR regulation is controlled by both GR itself and Erg direct interaction to GR gene (B).

Simulations of GR/Erg pathway in CEM C1–15 cells. The same process of simulation was carried out in Model 5 and Model 6 as described in Figure . Solid squares are the experimental data and the error bars are means ± SD for three sets of experiments. The black solid line represents the simulation by Model 5 and the blue solid line is the simulation of Model 6. (A) Protein time-course simulation of GR/Erg pathway in CEM-C1–15 cells. (B) GR and Erg mRNA time-course simulations in CEM-C1–15 cells.

Topology of GR/Erg models in CEM C1–15 cells. Schematic representation of GR/Erg pathway in CEM-C1–15 cells. The nature of the topologies was based on previously established direct and indirect GR target models in CEM-C1–15 cells (Chen et al., ). GR autoregulation was not included in CEM-C1–15, Model 5 and Model 6 differ by the GR regulation on Erg induction, with Model 5 indicating Erg as a direct GR target (A) and Model 6 showing Erg as an indirect GR target where de novo protein synthesis is required for Erg induction (B).

Simulations of GR/Erg pathway in CEM C7–14 cells. (A) Protein time-course simulation of GR/Erg pathway in CEM C7–14 cells. The same process of simulation was carried out in Model 3 and Model 4 as described in Figure . Solid squares are the experimental data and the error bars are means ± SD for three sets of experiments. The black solid line represents the simulation by Model 3 and the blue solid line is the simulation of Model 4. The residual value was calculated to assess the quality of the fit between the simulations and the experimental data. (B) GR and Erg mRNA time-course simulations in CEM C7–14 cells.

GR target gene protein expression in CEM cells. Western blot analysis of GR, Erg, c-Jun, and Bim protein levels, with actin as a control in C7 (A) and C1 (B) cells cultured with a combination of 1 μM Dex, 10 μM YK-4-279, 10 μM JNK inhibitor (SP600125) for 48 h. Protein levels were quantified by ImageJ, normalized to actin and presented as a histogram. Error bars represent ± SD of two independent experiments. An asterisk indicates a significant difference at p < 0.05 using ANOVA and Tukey’s test.

Simulations of GR/Jun/Bim induction in CEM C7–14 cells. The simulation process and the experimental data were described in our previous work (Chen et al., , ). In brief, the expression dynamics were simulated with the use of CellDesigner and SBML-PET parameter estimation tool based on the experimental data obtained at 0, 2, and 10 h after 1 μM dexamethasone (Dex) treatment (Chen et al., ). Solid squares are the mean of the normalized experimental data and bars are the SDs for three sets of experiments. The simulation process is divided to two steps, the parameters for GR activation alone were first obtained (dotted line), and then the rest of the parameters were estimated based on the individual model topologies. The black solid line represents the simulation by Model 1 and the blue solid line is the simulation of Model 2 (A) Protein time-course simulations of GR, c-Jun, and Bim in CEM C7–14 cells. (B) The simulations for GR, c-Jun, and Bim mRNA dynamics. The models as shown revealed the characteristic kinetics of GR, c-Jun, and Bim in response with Dex in CEM C7–14 cells. The residual value was calculated to assess the quality of the fit between the simulations and the experimental data.

Topology of GR/Jun/Bim models in CEM C7–14 cells Schematic representation of GR inducing Bim via c-Jun. The figure summarizes the basic mechanism of Bim regulation controlled by GR. The model topology was based on (Chen et al., , ), where glucocorticoid passes through the cell membrane, causes GR activation by dissociating GR from the cytoplasmic heat shock protein (HSP) complex. The bound GR dimerizes, either activates or represses its target genes through binding to GREs in the target genes or via the recruitment of other transcription factors. All models were constructed by CellDesigner, based on the known or potential molecular mechanisms but without taking the cytoplasmic-nuclear compartmentalization into account. Basal transcription, GR autoregulation, mRNA degradation, protein degradation, and binding dynamics were included in the models and the reactions were described using first order mass action kinetics. The details of the kinetic equations in all models are described in Appendix. (A) Model 1 represents GR induces Bim activation via direct binding to c-Jun. (B) Model 2 is similar to Model 1 but differs by the nature of the interaction between the GR and c-Jun. An extra step of protein synthesis was introduced for targeting down-stream target gene c-Jun in Model 2.