• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Theor Biol. Author manuscript; available in PMC Sep 7, 2008.
Published in final edited form as:
PMCID: PMC2080843
NIHMSID: NIHMS29652

Signaling Perturbations Induced by Invading H. pylori Proteins in the Host Epithelial Cells

A Mathematical Modeling Approach

Abstract

Helicobacter pylori (H. pylori), a gram-negative bacterium, infects the stomach of approximately 50% of the world population. H. pylori infection is a risk factor for developing chronic gastric ulcers and gastric cancer. The bacteria produce two main cytotoxic proteins: Vacuolating cytotoxin A (VacA) and Cytotoxin-Associated gene A (CagA). When these proteins enter the host cell they interfere with the host MAP Kinase and Apoptosis signaling pathways leading to aberrant cell growth and premature apoptosis. The present study expanded existing quantitative models of the MAP Kinase and Apoptosis signaling pathways to take into account the protein interactions across species using the CellDesigner tool. The resulting network contained hundreds of differential equations in which the coefficients for the biochemical rate constants were estimated from previously published studies. The effect of VacA and CagA on the function of this network were simulated by increasing levels of bacterial load. Simulations showed that increasing bacterial load affected the MAP Kinase signaling in a dose dependant manner. The introduction of CagA decreased the activation time of mapK signaling and extended activation indefinitely despite normal cellular activity to deactivate the protein. Introduction of VacA produced a similar response in the apoptosis pathway. Bacterial load activated both pathways even in the absence of external stimulation. Time course of emergence of transcription factors associated with cell division and cell death predicted by our simulation showed close agreement with that determined from a publicly accesible microarray dataset of H. pylori infected stomach epithelium. The quantitative model presented in this study lays the foundation for investigating the affects of single nucleotide polymorphisms (SNPs) on the efficiency of drug treatment.

Keywords: H. pylori, Computational model, mapK, EGF signaling, FasL signaling

1 Introduction

Helicobacter pylori (H. pylori) is a gram-negative bacterium that inhabits the acidic conditions of the human stomach in 50% of the world’s population [Hatakeyama and Brzozowski, 2006]. Chronic H. pylori infection is a major risk factor for gastric ulcers and gastric cancer [Hatakeyama and Brzozowski, 2006].

H. pylori effects host stomach epithelial cells by producing two unique cytotoxic proteins: Vacuolating cytotoxin A (VacA) and Cytotoxin-Associated gene A (CagA) [Cover and Blanke, 2005, Hatakeyama and Higashi, 2005]. In previously published experimental studies, VacA has been implicated in hyper-vacuolization, small molecule leakage and apoptosis [Cover and Blanke, 2005]. CagA has been implicated in morphological changes (“hummingbird” phenotype) and induction of the MAP Kinase pathway [Hatakeyama and Higashi, 2005].

VacA acts as both a trans-membrane channel and a protein activator [Galmiche et al., 2000, Cover et al., 2003]. VacA associates with planar lipid membranes, such as the outer cell membrane, vesicles and mitochondrial membranes [Galmiche et al., 2000]. When VacA associates with vesicle membranes it creates a leak channel, which eventually leads to vesicle swelling [Cover and Blanke, 2005]. This hyper-vacuolization is a drastic phenotypic effect noticed in vitro, however, rarely under in vivo conditions [Cover and Blanke, 2005]. Blocking the hyper-vacuolization phenotype has no effect on the VacA induced apoptosis rates, [Yamasaki et al., 2006] since the effect is independent of vesicle activity. When VacA associates with the mitochondrial membrane it begins to activate the BAX protein [Yamasaki et al., 2006]. This protein catalyzes the release of Cyt-C from the mitochondria which then begins the apoptosis cascade through the type II pathway downstream of the traditional activators [Clements et al., 2001].

CagA is injected into the host cytosol through a type IV secretory system [Bourzac and Guillemin, 2005]. Within the cytosol it can interact with the Grb-2 and SOS proteins to activate the MAP Kinase pathway in a Ras-dependent manner [Mimuro et al., 2002]. It is also phosphorylated by the host protein SRC Kinase [Fragale et al., 2004]. In the phosphorylated form, it can bind to the host protein Shp-2 and relieve its structural inhibition. Activated Shp-2 directly activates mapK in a Ras-independent manner [Tsutsumi et al., 2003]. CagA also activates the host protein CSK which then preferentially phosphorylates and deactivates SRC Kinase [Fragale et al., 2004], providing a feedback loop to modulate the activity of CagA.

MAP Kinase and Apoptosis pathways have been extensively studied and reviewed by others [Clarkson and Watson, 1999, Kohno and Pouyssegur, 2006, Lamkanfi et al., 2006]. Quantitative models of these pathways have been developed using rate equations and experimentally determined binding constants [Bhalla and Iyengar, 1995, Hua et al., 2005]. We used this body of research as a foundation to computationally simulate the interaction of H. pylori proteins with the host epithelial cell protein networks, in order to identify the critical modes of this interaction that lead to increasing rates of growth and apoptosis. We have compared our findings in a time dependent manner with the transcription factor activation data obtained from publicly accessible microarray data [El-Etr et al., 2004] and PAINT analysis [Vadigepalli et al., 2003].

Our results indicate a bacterial load dependent activation of the MAP Kinase and Apoptosis pathways. This dependency varies based on whether the pathway was stimulated by physiological conditions. The time dependent transcription factor activation predicted by the model equations agrees with the outcome of the microarray data, indicating that present mathematical formulations of MAP Kinase and Apoptosis pathways are able to reproduce response to bacterial challenges to the host cell.

2 Method

2.1 Quantitative Model

Quantitative interactions of H. pylori bacterial proteins with the host epithelial cell proteins are described below in a pathway specific manner. The pathways affected by bacterial proteins were previously shown to be the Map Kinase and the Apoptosis pathways in a large number of experimental studies as reviewed in [Censini et al., 2001, Cover et al., 2003]. Figure 1 shows a functional breakdown of the entire network model for visualization purposes.

Fig. 1
Diagram of the host signaling network (Map Kinase and Apoptosis pathways) perturbed by the invading bacterial proteins which are marked in red. The pathways are divided into functional groups for representative purposes. A diagram the entire network and ...

2.2 Map Kinase Signaling Pathway

The Map Kinase signaling pathway equations used in our study were previously presented by [Bhalla and Iyengar, 1995]. Their mathematical model contained 120 nodes (proteins in various phosphorylation states) and 200 reactions between the nodes. The rate equations used in the model consisted of simple association reactions and Michaelis-Menten enzyme kinetics. We used Locus Link ID comparison to confirm that the nodes in [Bhalla and Iyengar, 1995] were actually present in the KEGG Map Kinase pathway [Kanehisa and Goto, 2000]. Moreover we were able to confirm the interaction connections between these nodes of the Map Kinase pathway using the Ingenuity Pathway Analysis (IPA [www.ingenuity.com]). The rate equations governing the network nodes and their interactions with H. pylori proteins are presented in the supplementary information. The Map Kinase parameters and equations appearing in the supplementary information were previously published by [Bhalla and Iyengar, 1995] and more recently used by [Pant and Ghosh, 2005a,b].

The equations governing the interaction of the host signaling networks with the invading H. pylori proteins used in this study are described in Equations 1-7 in Table 1. In these equations the bacterial protein CagA interacts with GBR-2 and Shp-2 within the Map Kinase pathway [Censini et al., 2001] and also CagA is phosphorylated at EPIYA motifs by the host protein SRC Kinase [Naito et al., 2006]. The network diagram shown in Figure 1 indicates the need of prescribing 15 rate constants describing the interactions of CagA with the host. The two Michaelis-Menten rate constants, (k1, k2) from Equation 1 in Table 1, for the enzymatic phosphorylation of CagA by SRC Kinase were previously estimated by [Jeansonne et al., 2005] using affinity purification to isolate and characterize pure SRC Kinase.

Table 1
The equations that follow describe the direct interactions between host and bacterial proteins in the simulation. The terms in bold are taken directly from either Bhalla and Iyengar [1995] or Hua et al. [2005] The entire set of 200 differential equations ...

The forward and reverse binding constants, from Equation 2 in Table 1, involved in the interaction between phosphorylated CagA and c-SRC Kinase (CSK) (k3, k4) were estimated from data on immunoprecipitation and an in vitro kinase assay [Tsutsumi et al., 2003]. Previous studies show that active CSK phosphorylates and deactivates SRC Kinase, suggesting the presence of a feedback loop modulating the amount of active CagA [Hatakeyama and Higashi, 2005] as described in Equation 3 in Table 1. The rate constants for this interaction (k5, k6) were estimated based the activity of SRC Kinase as previously mentioned above [Jeansonne et al., 2005].

The next set of rate constants (k7, k8) involves the binding interaction between phosphorylated CagA and the host protein Shp-2 [Hatakeyama, 2002, Higashi et al., 2002, Zhou, 2002]. The constant k7 refers to the binding affinity between CagA and Shp-2 and k8 refers to the dissociation of the complex in Equation 4 in Table 1. Both k7 and k8 were estimated from on in-vitro kinetics on purified proteins [Hatakeyama, 2002]. Furthermore, Shp-2 activation produces a Ras-independent activation of mapK [Neel et al., 2003]. This leads to the next set of rate constants (k9, k10) for Equation 5 in Table 1 which describes the rate of activation of mapK by active Shp-2. In the absence of experimental data in the literature these constants were assumed to be the same as those regulating the activation by activated mapKK [Bhalla and Iyengar, 1995].

Non-phosphorylated CagA can also interact with the host GBR-2 protein [Mimuro et al., 2002] as per Equation 6 in Table 1. This interaction allows GBR-2 to recruit SOS and subsequently convert the inactive GDP-Ras to active GTP-Ras. GTP-Ras will then continue down the Map Kinase pathway in a Ras-dependent nature. The parameters (k11, k12) were estimated from experiments using in-vitro purified proteins by [Mimuro et al., 2002]. It is assumed that the CagA-SOS-GBR-2 complex activates Ras at the same rate as the SHC-SOS-GBR-2 complex [Bhalla and Iyengar, 1995].

2.3 Apoptosis Pathway

The Apoptosis Pathway equations used in this study were previously presented [Hua et al., 2005] and again we confirmed the constituency of its nodes with the corresponding KEGG pathway. Moreover an IPA search also identified the node connections specified. Overall this pathway has 56 nodes and 114 reaction equations. The Apoptosis pathway of the host epithelial cell is affected by the bacterial protein VacA as reviewed in [Cover and Blanke, 2005]. In this study it is assumed that VacA can enter the host cell cytoplasm through such mechanisms and focuses on its interaction with the Apoptosis pathway proteins. Estimates for the rate constants of BAX activation by VacA (k13, k14) for Equation 7 in Table 1 were derived from experiments on isolated mitochondria and purified cellular proteins [Galmiche et al., 2000].

Production of bacterial proteins CagA and VacA as represented by KprodXBL in the model equations were assumed to have occurred at a constant rate proportional to XBL. This variable is a measure of bacterial load (CagA and VacA) and corresponds to the Multiplicity of Infection (MOI) [Tsutsumi et al., 2003, Yamasaki et al., 2006].

2.4 Numerical Simulation

The mathematical model incorporating the interaction of invading bacterial proteins with the host Map Kinase and Apoptosis pathways was constructed with the CellDesigner tool [Funahashi et al., 2003]. This program creates an SBML [Hucka et al., 2003, Kitano et al., 2005] compliant model which can be simulated by a number of tools. The ordinary differential equations (ODEs) resulting from the model were solved using the MatLab 7.0.4 (www.mathworks.com) ode23s function with default values given for ode23s. The protein levels described by the initial conditions were kept constant in the simulations modeling the signal processing through the Map Kinase and Apoptosis pathways. The proteins in these pathways were allowed to change state as the during the transmission of the signal as defined by network equations.

The simulations take into account the possibility that host stimulation of the effected pathways modulate bacterial pathogenesis [Hatakeyama and Brzozowski, 2006]. To model host stimulation of the respective pathways, the host cells were assumed to be exposed to a concentration of EGF = 0.5 uM and FasL = 0.1 uM, resulting in subsequent stimulation of the Map Kinase [Bhalla and Iyengar, 1995] and Apoptosis pathways [Hua et al., 2005].

2.5 Parameter Scan

The parameters appearing in our simulated Map Kinase and pathways were investigated previously in a parameter search space by [Pant and Ghosh, 2005a,b]. This study focuses on the impact of parameters (k1-k14) governing the direct interaction of host and invading proteins. The standard values of these 14 parameters were multiplied by a modifying constant (Kmod) that ranged in value between 0.1 and 4.0. Simulations were then executed by altering these parameters one at a time and quantifying the time course of mapK and Casp-3 activity. Note that Casp-3 stands as a phenotypic marker for apoptosis effects and mapK for growth.

2.6 Microarray analysis

Publicly accessible microarray datasets involving global gene expression profiles of human epithelial cells in the presence and absence of H. pylori infection were acquired from Stanford Source Microarray Database [Gollub et al., 2006]. Specifically, these data sets corresponded to microarray data on T84 polarized monolayers taken at discrete time intervals following infection with wild-type H. pylori and their controls. The Stanford gene chip used in the microarray study consisted of cDNA spots for approximately 43,000 genes [El-Etr et al., 2004]. The data sets were normalized based on the standard Stanford procedure [Gollub et al., 2006] into log2 normalized data. Table 2 summarizes the microarray data used in this study.

Table 2
A summary of microarray data from El-Etr et al. used in the analysis.

Microarray data samples from infected cells were grouped and compared to the entire time-course of the corresponding data for mock infected cells (control data). This analysis was performed using the t-test module of the TIGR MEV tool [Welch, 1947, Dudoit et al., 2002, Pan, 2002]. Genes which had a p-value less than 0.01 regardless of the direction of expression change were saved for further analysis. The pairings used in the analysis are listed in Table 2.

Since both the Map Kinase and Apoptosis pathways effect transcription of functional genes the gene sets should have an over-representation of promoters for either apoptosis or growth. The PAINT promoter tool [Vadigepalli et al., 2003] is a high-throughput bioinformatics algorithm for promoter analysis. The PAINT analysis retrieved the 2000 bases upstream of each gene within the test set and searched the public TRANSFAC database [Matys et al., 2003] for known promoter recognition sites and their corresponding transcription factors.

3 Results and Discussion

This section presents the simulation results concerning perturbations of the Apoptosis and MAP Kinase pathways, which are due to H. pylori infection. The parameters that govern the interaction of bacterial proteins with the host protein network are altered one at a time to gage the possible effects of bacterial and host genotypes on the predicted results. The last subsection compares model predictions with findings based on microarray data of epithelial cells infected with H. pylori and their controls.

3.1 MAP Kinase pathway

The bacterial protein CagA interacts with the proteins of the host cell, and these proteins lay in the MAP Kinase Pathway as described above. CagA can directly activate mapK through activation of Shp-2 or activate RAS through interaction with GBR-2 and SOS. Consistent with the extent of interaction of CagA with the host protein network, our results indicate a complex set of perturbations to the MAP Kinase pathway, as quantified in simulations by the time variant concentration of active mapK (Figure 2). The simulations, shown in Figure 2, were generated by solving the MAP Kinase pathway rate equations for a time-span of 32400 second (9 hours) under EGF stimulated (A) and non-stimulated (B) conditions. Note that EGF (Epidermal Growth Factor) regulates the growth activity of many types of epithelial cells, including those affected by H. pylori infection. The figure indicates that the level of activation of mapK changes with bacterial load in a dose dependent manner. Furthermore, bacterial load shortens the time to mapK activation in a dose dependent manner. The time of activation of MAP Kinase pathway due to bacterial infection is shorter in the presence of EGF stimulation at moderate to low bacterial loads, but time to activation decreases rapidly with increasing bacterial load as shown in Figure 2 C. In contrast, the time for deactivation is independent of the level of bacterial load.

Fig. 2
The time course of active mapK due to bacterial loading in the presence and absence of EGF. The symbol BL refers to the simulated level of bacterial load. A: The system was stimulated by a constant EGF concentration (EGF = 0.5 uM) and varying bacterial ...

The time course of MAP Kinase activation due H. pylori infection can be better understood when the simulation results are plotted as a function of time for the critical nodes appearing within the pathway (Figure 3). The MAP Kinase nodes of the figure include GBR-2-SOS-SHC complex, GTP-RAS, active Raf, RAS-Raf complex, active mapKK, active mapK, CagA-Shp-2 complex, inactivated GBR-2-SOS and aberrantly activated GBR-2-SOS-CagA. The time variant concentration curves of the nodes within the MAP Kinase pathway under EGF stimulation are shown as a black line with stars at specific data points; corresponding curves due to bacterial loading only are shown as a red line with circles at specific data points; the same curves with both bacterial loading and EGF stimulation conditions are shown as a blue line with triangles at specific data points. The results shown in the diagram point to a feedback loop of active mapK deactivating the GBR-2-SOS complex and modulating the activation length of the MAP Kinase pathway. MapK activated by bacterial proteins causes an aberrant feedback loop, which deactivates the GBR-2-SOS complex and prevents the other regulatory functions of the MAP Kinase pathway from deactivating mapK.

Fig. 3
A diagram indicating the simulation of the Map Kinase pathway under stimulation by EGF only (black starred line), Bacterial load only (red circled line) and both (blue triangled line). Bacterial load was held at 0.1 for the simulation. Each node represents ...

3.2 Apoptosis pathway

The interaction of bacterial protein VacA with the BAX protein of the Apoptosis pathway leads to a perturbation of the Apoptosis network, as shown in Figure 1. The end result of this perturbation is exhibited in the concentration of the active form of Casp-3, which is considered to be a marker for upcoming apoptosis [Xia et al., 2001]. Our results, obtained by solving the 114 rate equations of the apoptosis model for a time span of 86400 seconds (24 hours), show that Casp-3 activation is dependent on the presence of FasL, an Apoptosis stimulating hormone, stimulation of the host cell and the bacterial load, as measured by the concentration of VacA in the host cell (Figure 4A). FasL concentration in this computational simulation was held constant at 0.1uM (I) or zero (II), which is consistent with standard experimental techniques [Kelly et al., 2002]. Bacterial load was then varied for both stimulated and non-stimulated conditions from 0.1 to 1.0. Our results indicate that FasL activation is much more e cient in a time-wise manner for inducing apoptosis compared to H. pylori infection in short time periods (7 hours vs. > 2 hours), but bacterial infection is sufficient to cause apoptosis by itself within the first twenty-four hours of infection.

Fig. 4
The time course of active Casp-3 under bacterial loading in the presence and absence of FasL stimulation. The symbol BL refers to the simulated level of bacterial load. A I: The system was stimulated by a constant FasL concentration FasL = 0.1uM and varying ...

Infection induced apoptosis is dependent on the presence FasL stimulation of the host epithelial cells, which might have been caused by a number of factors including inflammation. Figure 4B shows that FasL induced, Casp-3 activation proceeds independent of the level of bacterial infection, whereas, the bacterial infection induced apoptosis is highly dependent on the cellular concentration of the invading VacA protein. The 80% activation time, shown in the figure, identifies the time in which 80% of Casp-3 had been converted to its active form. These results point to the importance of bacterial load in the progression H. pylori induced gastritis.

The model simulations presented here bring a deeper understanding of the detailed interrelations between various nodes within the Apoptosis pathway. The diagram in Figure 5 shows the predicted time variant concentrations of critical nodes within the Apoptosis pathway under FasL stimulated and bacterial loading conditions. Each curve in a node box in Figure 5 represents the concentration of the node under bacterial load only (red with circles), stimulated only (black with stars) and stimulated and bacterial loading (blue with triangles) conditions. The numerical simulation indicates two distinct modes of flow of the apoptosis inducing signal through the pathway. The preferred mode of flow for FasL stimulation is through activation of Casp-3 directly by Casp-8, whereas, the bacterial load induced signal perpetuates through an alternative cycle that begins with BAX activation by VacA and goes through Cyt-C release and subsequent activation of Casp-3. These results are in agreement with medical practice of eradicating H. pylori infections to prevent progression from gastric ulcers to cancerous regions [Xia et al., 2001].

Fig. 5
Simulation of the Apoptosis pathway under FasL stimulation (black starred line), bacterial only (red circled line) and both (blue triangled line). Bacterial load was held at 0.1 for the simulation. Each node represents the time-dependant concentration ...

Activation of the Apoptosis pathway by H. pylori infection bypasses the primary mode of transmission that utilizes the interaction between Casp-8 and Casp-3. Under bacterial loading only conditions Casp-8 never activates and interactions 2, 7, and 8 never occur, since VacA perturbations activate the pathway downstream of Casp-8, as shown in Figure 5. The VacA induced activation of BAX increases Cyt-C release through interaction 4. This causes a slower rate Casp-9 activation, as compared to FasL stimulation, due to the interaction of XAIP, which inhibits Casp-9 activation (not shown). Active Casp-9 then activates Casp-3 through interactions 6 and 9. The results of this simulation demonstrate that bacterial induced apoptosis bypasses the primary cellular modes of apoptosis, avoiding various checkpoints for cell death.

When the simulated host cell is exposed to both FasL stimulation and bacterial loading conditions, the FasL mode of transmission dominates. Figure 5, FasL stimulation leads to activation of Casp-8 and subsequent activation of Casp-3 after approximately three hours. Even under high bacterial loads the Casp-3 activation by Casp-8 preempts the hyper-activation of Casp-9 by bacterial induced Cyt-C release.

3.3 Bacterial Parameters and Their Effect of Host Cell Response

Parameters that govern the interaction of bacterial proteins with the host cell proteins were described above. We have changed these parameters one at a time by multiplying their standard values with a following coefficients (0.1, 0.2, 0.5 1, 2, 4). Results of this large scale simulation show that 12 out of the 14 parameters described above could be varied by large amounts without significant alteration to the time course of mapK activation. On the other hand, as shown in Figure 6, the parameter k9, representing the rate at which the CagA-Shp-2 complex converts inactive mapK into activated mapK (Figure 3, Interaction 6), strongly influences time to activation of mapK. Similarly, the decreasing values of the rate constant k8, governing the dissociation rate of the CagA-Shp-2 complex, shortened the activation time of mapK. These results are consistent with medical data, indicating genotype dependence of H. pylori as a risk factor for infection induced gastritis [Bhat et al., 2005]. Moreover, our simulation results point to the importance of single nucleotide polymorphisms that affect the formation and enzymatic efficiency of Shp-2.

Fig. 6
The top panel shows time course curves of mapK activation due to variation in k9 and the bottom panel indicates the time course of mapK due to variation in k8. Simulation was performed under non-stimulated conditions with a BL=1.

3.4 Comparison of Simulation Results to Microarray Data

The end result of signal transduction through the MAP Kinase pathway is the emergence of a series of transcription factors that modulate the expression of growth related genes. The surrogate markers for the transcription factors involved in our simulation are mapK and PKA. Increasing concentrations of these nodes point to increasing expression of genes related to the transcription factors such as AP-1 and elk-1. Similarly, the end result of apoptosis pathway is Casp-3 activation, which leads to a sequence of events such as activation of the transcription factors Oct-1 and NFkB. Using PAINT and IPA tools and publicly accessible microarray data on H. pylori infected epithelial cells and their controls [El-Etr et al., 2004], we determined the time course of the emergence of transcription factors that are linked to the concentration of MapK, PKA, and Casp-3, as shown in Table 3. The impetus for this study was to see if the time course of emergence of MAP Kinase and Apoptosis related transcription factors matched with the predicted time course of signal transduction presented in this study.

Table 3
Microarray data indicating the time course of emergence of transcription factors (TF) due to H. pylori infection. The table shows the results of the promoter analysis of the paired time points vs. the entire mock time-course using the PAINT software (p<0.1). ...

Upon submitting to PAINT the significant genes found in the analysis of the 2 and 4 hour time points, PAINT returned a list of transcription factors which showed a statistical over-representation; these are listed in Table 3. The transcription factors that appear at this early stage of infection include the mapK activated elk-1 (p=0.00000) and AP-1 (p=0.00000), a transcription known to be activated by mapKK [Takeuchi et al., 2006]. These results closely match with the time course of signaling of the MAP Kinase pathway predicted in this study (Figure 7).

Fig. 7
Normalized activities of selected proteins which have been associated with the AP-1, Elk-1, NfKb, Oct-1 transcription faction. The simulation was integrated for 86400 seconds (24 hours) with a BL = 1.0 and (FasL = 0.1 uM, EGF = 0.5 uM).

The transcription factor NFkB appeared as significantly over-represented (p=0.0083) in microarray data at the 8 and 12 hour comparison. This transcription factor controls survival and inflammatory genes and is activated in response to apoptosis [Clarkson and Watson, 1999]. Recent studies have suggested that NFkB can be activated by Casp-8 [Lamkanfi et al., 2006]. Figure 7 shows the Casp-8 response to infection and stimulation. Oct-1 was also significantly over expressed at 12 to 24 hour time points; this transcription factor is known to be down-regulated by PKA activity and activated in response to DNA damage [Heckman et al., 2006]. During the early time-points, Oct-1 activity is repressed by high levels of PKA activity (Figure 7). However, active Casp-3 will begin apoptotic events and increase DNA damage [Lamkanfi et al., 2006]. Again, these results match with the predicted time course of Apoptosis signaling events using the CellDesigner simulation, as shown in Figure 7.

4 Conclusion

This study integrates detailed analysis of signaling pathways related to H. pylori infection of epithelial cells with global gene expression patterns that are associated with infection and the time course of the emergence of transcription factors which modulate the growth and death of infected epithelial cells. The signaling pathway models used in this study are composed of sets of ODEs with time as an independent variable and concentrations of proteins appearing in these signaling pathways as dependent variables. These equations and the estimates of rate parameters associated with them were previously published. We have incorporated a set of new rate equations and associated rate constants to these equations in order to quantitatively describe the binding interactions of H. pylori proteins with the protein network of human epithelial cells. One of the most important findings of our study is the fact that existing mathematical models of MAP Kinase and Apoptosis pathways stood up to the challenge of comparison with microarray data on H. pylori. Note that the signaling pathways investigated in this study are implicated in the transformation of human cells to a cancer phenotype.

Our study showed the importance of the feedback loop in MAP Kinase mediated signaling as a modulator of bacterial activation of the pathway. Our simulations show that bacterial load activates the MAP Kinase pathway and subsequently a feedback loop of mapK; specifically, the feedback loop operates by deactivating the RAS-RAF complex and results in aberrant mapK signaling. The two constants that showed significant effect on the time course of simulated infection are linked to the binding of Shp-2 to CagA and the Shp-2 mis-activation of mapK. These proteins might be the loci of SNPs which modulate host response and the time course of infection. On the bacterial side CagA SNPs have already been shown to have drastic effects on the host response [Israel et al., 2001, Höcker and Hohenberger, 2003, Naito et al., 2006].

The transcription factors that emerge due to the MAP Kinase signaling perturbations by bacterial infection lead to an increased transcription of Apoptosis inhibitors XAIP and FLIP [Kasid and Dritschilo, 2003, Shelton et al., 2003]. In addition, the bacterial protein VacA specifically binds and activates BAX, a protein that appears in the FasL stimulated Apoptosis pathway. Our simulations concerning the time course of Apoptosis signaling are consistent with aforementioned mentioned microarray data on H. pylori infected epithelial cells and their controls. Our simulations indicated that SNPs in the BAX protein are likely to affect the interactions of the VacA protein and therefore, the time course of infection. Genotyping of patients with H. pylori infection are only beginning. The use of recently developed high-throughput genotyping tools will lead to testing and validating the findings of this study concerning the SNPs in the human BAX, Shp-2 and mapK proteins.

Supplementary Material

01

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • Bhalla US, Iyengar R. Emergent Properties of Networks of Biological Signaling Pathways. Chem. Biol. 1995;2:489.
  • Bhat N, Gaensbauer J, Peek RM, Bloch K, Tham KT, Blaser MJ, Perez-Perez G. Local and Systemic Immune and Inflammatory Responses to Helicobacter pylori Strains. Clinical and Vaccine Immunology. 2005;12(12):1393–1400. [PMC free article] [PubMed]
  • Bourzac KM, Guillemin K. Helicobacter pylori-host cell interactions mediated by type IV secretion. Cellular Microbiology. 2005;7(7):911–919. [PubMed]
  • Censini S, Stein M, Covacci A. Cellular responses induced after contact with Helicobacter pylori. Curr Opin Microbiol. 2001;4(1):41–6. [PubMed]
  • Clarkson RWE, Watson CJ. NF-kappaB and Apoptosis in Mammary Epithelial Cells. Journal of Mammary Gland Biology and Neoplasia. 1999;4(2):165–175. [PubMed]
  • Clements JD, Kuck D, Kolmerer B, Iking-Konert C, Krammer PH, Stremmel W, Rudi J. Vacuolating Cytotoxin of Helicobacter pylori Induces Apoptosis in the Human Gastric Epithelial Cell Line AGS. Infection and Immunity. 2001;69(8):5080–5087. [PMC free article] [PubMed]
  • Cover TL, Blanke SR. Helicobacter pylori VacA, a paradigm for toxin multifunctionality. Nat Rev Microbiol. 2005;3(4):320–32. [PubMed]
  • Cover TL, Krishna US, Israel DA, Peek RM. Induction of Gastric Epithelial Cell Apoptosis by Helicobacter pylori Vacuolating Cytotoxin. 2003;1 [PubMed]
  • Dudoit S, Yang YH, Callow MJ, Speed TP. Statistical methods for identifying differentially expressed genes in replicated cDNA microarray experiments. Statistica Sinica. 2002;12(1):111–139.
  • El-Etr SH, Mueller A, Tompkins LS, Falkow S, Merrell DS. Phosphorylation-independent effects of CagA during interaction between Helicobacter pylori and T84 polarized monolayers. J Infect Dis. 2004;190(8):1516–23. [PubMed]
  • Fragale A, Tartaglia M, Wu J, Gelb BD. Noonan syndrome-associated SHP 2/PTPN 11 mutants cause EGF-dependent prolonged GAB 1 binding and sustained ERK 2/MAPK 1 activation. Human Mutation. 2004;23(3):267–277. [PubMed]
  • Funahashi A, Morohashi M, Kitano H. CellDesigner: a process diagram editor for gene-regulatory and biochemical networks. BioSilico. 2003;1:159–162.
  • Galmiche A, Rassow J, Doye A, Cagnol S, Chambard JC, Contamin S, de Thillot V, Just I, Ricci V, Solcia E, et al. The N-terminal 34 kDa fragment of Helicobacter. The EMBO Journal. 2000;19(23):6361–6370. [PMC free article] [PubMed]
  • Gollub J, Ball CA, Sherlock G. The Stanford Microarray Database: a user’s guide. Methods Mol Biol. 2006;338:191–208. [PubMed]
  • Hatakeyama M. Deregulation of SHP-2 tyrosine phosphatase by the Helicobacter pylori virulence factor CagA. Keio J Med. 2002;51(2):26–32. [PubMed]
  • Hatakeyama M, Brzozowski T. Pathogenesis of Helicobacter pylori Infection. Helicobacter. 2006;11(1):14–20. [PubMed]
  • Hatakeyama M, Higashi H. Helicobacter pylori CagA: a new paradigm for bacterial carcinogenesis. Cancer Sci. 2005;96(12):835–843. [PubMed]
  • Heckman CA, Duan H, Garcia PB, Boxer LM. Oct transcription factors mediate t (14; 18) lymphoma cell survival by directly regulating bcl-2 expression. Oncogene. 2006;25:888–898. [PubMed]
  • Higashi H, Tsutsumi R, Muto S, Sugiyama T, Azuma T, Asaka M, Hatakeyama M. SHP-2 Tyrosine Phosphatase as an Intracellular Target of Helicobacter pylori CagA Protein. Science’s STKE. 2002;295(5555):683–686. [PubMed]
  • Höcker M, Hohenberger P. Helicobacter pylori virulence factorsone part of a big picture. The Lancet. 2003;362(9391):1231–1233. [PubMed]
  • Hua F, Cornejo MG, Cardone MH, Stokes CL, Lauffenburger DA. Effects of Bcl-2 Levels on Fas Signaling-Induced Caspase-3 Activation: Molecular Genetic Tests of Computational Model Predictions1. The Journal of Immunology. 2005;175(2):985–995. [PubMed]
  • Hucka M, Finney A, Sauro HM, Bolouri H, Doyle JC, Kitano H, Arkin AP, Bornstein BJ, Bray D, Cornish-Bowden A, Cuellar AA, Dronov S, Gilles ED, Ginkel M, Gor V, II, Goryanin WJ, Hedley TC, Hodgman JH, Hofmeyr PJ, Hunter NS, Juty JL, Kasberger A, Kremling U, Kummer N, Le Novere L, Loew M, Lucio D, Mendes P, Minch E, Mjolsness ED, Nakayama Y, Nelson MR, Nielsen PF, Sakurada T, Scha JC, Shapiro BE, Shimizu TS, Spence HD, Stelling J, Takahashi K, Tomita M, Wagner J, Wang J. The systems biology markup language (sbml): a medium for representation and exchange of biochemical network models. Bioinformatics. 2003;19(4):524–531. [PubMed]
  • Israel DA, Salama N, Arnold CN, Moss SF, Ando T, Wirth HP, Tham KT, Camorlinga M, Blaser MJ, Falkow S, et al. Helicobacter pylori strain-specific differences in genetic content, identified by microarray, influence host inflammatory responses. Journal of Clinical Investigation. 2001;107(5):611–620. [PMC free article] [PubMed]
  • Jeansonne DP, Bordes TJ, Bennett CA, Kothandaraman G, Bush JG, Vaccaro JA. A rapid ATP affinity-based purification for the human non-receptor tyrosine kinase c-Src. Protein Expr Purif. 2005 [PubMed]
  • Kanehisa M, Goto S. KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Research. 2000;28(1):27–30. [PMC free article] [PubMed]
  • Kasid U, Dritschilo A. RAF antisense oligonucleotide as a tumor radiosensitizer. Oncogene. 2003;22:5876–5884. [PubMed]
  • Kelly E, Won A, Refaeli Y, Van Parijs L. IL-2 and Related Cytokines Can Promote T Cell Survival by Activating AKT1. The Journal of Immunology. 2002;168(2):597–603. [PubMed]
  • Kitano H, Funahashi A, Matsuoka Y, Oda K. Using process diagrams for the graphical representation of biological networks. Nature Biotechnology. 2005;23:961–966. [PubMed]
  • Kohno M, Pouyssegur J. Targeting the ERK signaling pathway in cancer therapy. Annals of Medicine. 2006;38(3):200–211. [PubMed]
  • Lamkanfi M, Declercq W, et al. Caspases leave the beaten track: caspase-mediated activation of nf-kappab. Journal of Cell Bioliology. 2006;2(173):165–171. [PMC free article] [PubMed]
  • Matys V, Fricke E, Geffers R, Gößling E, Haubrock M, Hehl R, Hornischer K, Karas D, Kel AE, Kel-Margoulis OV, et al. TRANSFAC® R: transcriptional regulation, from patterns to profiles. Nucleic Acids Research. 2003;31(1):374–378. [PMC free article] [PubMed]
  • Mimuro H, Suzuki T, Tanaka J, Asahi M, Haas R, Sasakawa C. Grb 2 Is a Key Mediator of Helicobacter pylori CagA Protein Activities. Molecular Cell. 2002;10(4):745–755. [PubMed]
  • Naito M, Yamazaki T, Tsutsumi R, Higashi H, Onoe K, Yamazaki S, Azuma T, Hatakeyama M. Influence of EPIYA-repeat polymorphism on the phosphorylation-dependent biological activity of Helicobacter pylori CagA. Gastroenterology. 2006;130(4):1181–90. [PubMed]
  • Neel BG, Gu H, Pao L. The’Shp’ing news: SH2 domain-containing tyrosine phosphatases in cell signaling. Trends Biochem Sci. 2003;28(6):284–93. [PubMed]
  • Pan W. A comparative review of statistical methods for discovering differentially expressed genes in replicated microarray experiments. Bioinformatics. 2002;18(4):546–554. [PubMed]
  • Pant D, Ghosh A. Automated oncogene detection in complex protein networks with applications to the MAPK signal transduction pathway. Biophysical Chemistry. 2005a;113:275–288. [PubMed]
  • Pant DK, Ghosh A. A systems biology approach for the study of cumulative oncogenes with applications to the MAPK signal transduction pathway. Biophys Chem. 2005b [PubMed]
  • Shelton JG, Steelman LS, Lee JT, Knapp SL, Blalock WL, Moye PW, Franklin RA, Pohnert SC, Mirza AM, McMahon M, et al. Effects of the RAF/MEK/ERK and PI3K/AKT signal transduction pathways on the abrogation of cytokine-dependence and prevention of apoptosis in hematopoietic cells. Oncogene. 2003;22:2478–2492. [PubMed]
  • Takeuchi K, Motoda Y, Ito F. Role of transcription factor activator protein 1 (AP1) in epidermal growth factor-mediated protection against apoptosis induced by a DNA-damaging agent. FEBS Journal. 2006;273:3743. [PubMed]
  • Tsutsumi R, Higashi H, Higuchi M, Okada M, Hatakeyama M. Attenuation of helicobacter pylori caga x shp-2 signaling by interaction between caga and c-terminal src kinase. J Biol Chem. 2003;278:3664–3670. [PubMed]
  • Vadigepalli R, Chakravarthula P, Zak DE, Schwaber JS, Gonye GE. PAINT: A Promoter Analysis and Interaction Network Generation Tool for Gene Regulatory Network Identification. Omics A Journal of Integrative Biology. 2003;7(3):235–252. [PubMed]
  • Welch BL. The Generalization of Student’s’ Problem when Several Different Population Variances are Involved Biometrika 1947. 341228–35.35www.ingenuity.com Ingenuity pathway analysis [PubMed]
  • Xia HHX, MM MB, Talley NJ. Apoptosis in gastric epithelium induced by helicobacter pylori infection: implications in gastric carcinogenesis. Apoptosis. 2001;96(1):16. [PubMed]
  • Yamasaki E, Wada A, Kumatori A, Nakagawa I, Funao J, Nakayama M, Hisatsune J, Kimura M, Moss J, Hirayama T. Helicobacter pylori Vacuolating Cytotoxin Induces Activation of the Proapoptotic Proteins Bax and Bak, Leading to Cytochrome c Release and Cell Death, Independent of Vacuolation. Journal of Biological Chemistry. 2006;281(16):11250. [PubMed]
  • Zhou X. SHP-2 tyrosine phosphatase-the target of CagA. Trends Microbiol. 2002;10(4):164. [PubMed]
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...