NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Madame Curie Bioscience Database [Internet]. Austin (TX): Landes Bioscience; 2000-2013.

Cover of Madame Curie Bioscience Database

Madame Curie Bioscience Database [Internet].

Show details

Decoding the Signaling Mechanism of Toll-Like Receptor 4 Pathways in Wild Type and Knockouts


* Corresponding Author: Bioinformatics Institute, A*STAR, Biopolis, Singapore 138-671 Institute for Advanced Biosciences, Keio University, Tsuruoka, 997-0035, Japan.

The Myeloid Differentiation Primary-Response Protein 88 (MyD88)-dependent and—independent pathways induce proinflammatory cytokines when toll-like receptor 4 (TLR4) is activated through lipopolysaccharide (LPS) stimulus. Recent studies have implicated a crosstalk mechanism between the two pathways. However, the exact location and nature of this interaction is poorly understood. Using my previous ordinary differential equations-based computational model of the TLR4 pathway, I investigated the roles played by the various proposed crosstalk mechanisms by comparing in silico nuclear factor κB (NF-κB) and Mitogen-Activated Protein (MAP) kinases dynamic activity profiles with experimental results under various conditions in macrophages to LPS stimulus (MyD88 deficient, TRAF-6 deficient etc.). The model that best represents the experimental findings suggests that the pathways interact at more than one location: i) TRIF to TRAF-6, ii) TRIF-RIP1-IKK complex and iii) TRIF to cRel via TBK1.


The Toll-like receptors (TLRs) are key elements of the innate immune system. These receptors recognize conserved pathogen-associated molecular patterns related to microorganisms, such as lipopolysaccharide (LPS) and double-stranded RNA and trigger both microbial clearance and the induction of immunoregulatory chemokines and cytokines. There are a total of 13 known TLRs to date, of which TLR4 has received particular attention.1,2 Upon LPS ligation, TLR4 activates the MyD88-dependent and MyD88-independent pathways. The MyD88-dependent pathway, which is common to all TLRs except TLR3, activates NF-κB and activator protein-1 (AP-1) resulting in the induction of proinflammatory chemokines and cytokines such as Tumour-Necrosis-Factor α (TNF-α) and interleukin-1β (IL-1β). The MyD88-independent pathway, on the other hand, activates Interferon (IFN) Regulatory Factor 3 (IRF-3) and induces IFN- β and other chemokines like CCL5 and CXCL10.1,2

Studies of signaling cascades mediated through the MyD88-dependent and -independent pathways have so far been predominantly performed in a nonconstitutive manner. That is, the two pathways have been studied independently of each other. More recently there have been implications that the components of the two pathways may indeed interact downstream of TRIF and, therefore, may be dependent on each other in the activation of transcription factor NF-κB. For example, the interaction of TRIF with TNF-Receptor-Associated Factor 6 (TRAF6) has been suggested by Sato et al (2003).3 This leads to the question whether TRAF6 binding to TRIF could lead to the activation of NF-κB in a MyD88-independent manner. However, other studies involving the LPS-induced activation of TLR4 in TRAF-6 deficient mice have also shown the induction of NF-κB.4,5 Collectively, these studies indicate a possible link between the MyD88-dependent and -independent pathways in the activation of NF-κB in both MyD88 and TRAF-6 deficient mice. However, the suggestion of signaling crosstalk occurring by the binding of TRAF-6 to TRIF seems controversial. 3,5

I approached this issue in a systemic manner. Previously, I developed a computational model of the MyD88-dependent and -independent pathways.6 The model showed the possible signaling mechanism for the delayed NF-κB kinetics observed in MyD88-deficient mice. Using this model with careful modifications, I investigated several in silico crosstalk mechanisms between the MyD88-dependent and -independent pathways and compared the model simulations with experimental findings for NF-κB and MAP kinases activity in wild type and various knock-out conditions.

Materials and Methods

The details of the modeling strategy and the original computational model (reference model) have been previously published.6 In short, the development of our model includes selecting appropriate signaling reaction networks and determining associated kinetic parameters. As the TLR field is relatively new, we do not know the kinetic details of each signaling process. In addition, although biological networks in general can behave in nonlinear fashion, in my original model I and also others, showed that downstream signaling reaction events to receptor activation can be described by first order mass action kinetics.7,8 Therefore, in this paper all signaling reactions are described by ordinary differential equations of first order.

Using existing knowledge of the TLR4 pathway (Fig. 1A, reference pathway) and mass-action kinetics, I chose the parameter values to fit wild-type (WT) semi-quantitative profiles of NF-κB and JNK activity.9 I next tested whether the same model parameters also performed well in a knockout (KO) condition, namely MyD88 KO condition.9 This is an iterative process where parameter values are selected using semi-quantitative NF-κB and JNK activity profiles in both WT and MyD88 KO conditions.6

Figure 1A. See following page for figure legend.

Figure 1A

See following page for figure legend.

My model begins with the TLR4 receptor in an active state through the binding of LPS. The active signal triggers both the MyD88-dependent and -independent pathways. For the MyD88-dependent pathway (Fig. 1A, reference model), the signaling reactions are (i) MyD88/ MAL associates to the TLR4 receptor (TIRAP can be lumped with MyD88), (ii) IRAK1 and IRAK4 associate with MyD88 at the receptor, (iii) the IRAK-MyD88 complex activates TRAF6, (iv) TRAF6 stimulates the formation of a TAB1/TAB2/TAK1 complex, (v) the TAB1/TAB2/ TAK1 complex triggers MKK3/6, MKK4/7 and IKK complexes (IKKα, IKKβ and IKKγ), (vi) MKK4/7 activates JNK (vii) MKK3/6 activates p38, (vii) IKKs phosphorylate IκBα and release NF-κB, (viii) p38 and JNK translocate to the nucleus, (ix) NF-κB translocates to the nucleus, (x) JNK and p38 activate AP-1, and (xi) NF-κB and AP-1 bind to the relevant gene promoters and induces transcription.

The following constitutes the MyD88-independent pathway: a) TLR4 stimulates intermediate 1 (I1), I2 and I3, b) I3 activates TRAM, c) TRIF is recruited to the TIR domain of TLR4 together with TRAM, d) TRIF binds TBK1 and activates IRF-3, e) TBK1 also activates cRel of NF-αB and f ) IRF-3 and NF-κB translocate to the nucleus and induce the relevant gene transcription.

Although I simulate quantitative results of the various activated proteins and protein complexes in response to TLR4 activation, I only make semi-quantitative comparisons between the simulation results and the experimental findings. This is due to the general lack of quantitative experimental data. In addition, I restricted my model simulations to 60 min after LPS stimulation, to ignore secondary signaling such as autocrine TNF-κ signaling and IκBα negative feedback regulation. I assume such secondary complexities are negligible within the time frame of my analysis.5,10

The initial conditions in the model are that apart from the signaling step TLR4 to MyD88/ MAL and I1, all other signaling processes begin with null activation (at t = 0). The various KO conditions were generated from the wild-type model by setting the reaction(s) upstream of the molecules to be null. I compared the simulation of JNK, p38 and NF-κB activity with published data and progressed from reference model to Model A to Model B, the latter being the most representative of the TLR4 signaling pathway.

All models were constructed and solved using E-Cell version 3.11 The complete computational wild-type Model B with the kinetic expressions and parameters is available upon request (


TRAF-6 Independent NF-κB Activation in MyD88 KO Is Possible

To test the proposed mechanism of crosstalk between TRAF-6 and TRIF that leads to the NF-κB activation in the MyD88-independent manner,3 I inserted, in silico, a signaling reaction between TRIF and TRAF-6 in the original reference model and labeled the updated model as Model A (Fig. 1B and Appendix A). Using this model, I simulated the NF-κB activity profile for WT and MyD88 KO conditions, initially setting the TRIF to TRAF-6 reaction null, and checked whether the model simulation mimics the experimental observation of Kawai et al, 1999 (Fig. 1C, WT & MyD88 KO).

It is well known that for wild type macrophages the MyD88-dependent pathway is the key pathway for early phase NF-κB activation. Therefore, to investigate the importance of TRIF and TRAF-6 crosstalk, I performed the NF-κB simulation at various rates of reaction between TRIF and TRAF-6 in MyD88 KO conditions (Fig. 1C, MyD88 KO- CR1 etc.). Interestingly, we observed that as the rate constant between TRIF-TRAF-6 is increased, the peak levels of NF-κB activity approaches the WT profile, even in MyD88 KO conditions, although with a time-delay response. This implies that TRIF to TRAF-6 crosstalk can result in MyD88-independent activation of NF-κB; however, it may not be dominant.

MyD88-Independent Pathway also Interacts Downstream of TRAF-6 but Upstream of TBK1

In TRAF6-deficient mice, delayed activation of MAP kinases has been reported.5 This observation indicates that in addition to TRIF to TRAF-6 and TRIF to cRel via TBK1 (Fig. 1B) there exists other crosstalk mechanisms, intuitively, upstream of MAP kinases and downstream of TRAF-6. I next tested my model in silico by making interactions between TRIF and signaling molecules/complexes downstream of TRAF-6 but upstream of JNK. This reduced the possibility of crosstalk of TRIF with either the TAK/TAB complex or MKK3/6 and MKK4/7 (Fig. 1B).

Recently, Transforming-Growth-Factor-β-Activated Kinase (TAK1) has been shown to play a vital role in multiple signaling pathways.12 TAK1 KO studies revealed that TAK1 is essential for the TLR-induced JNK activity12 and the activation of MAP kinases in response to IL-1β.13 To test whether TRIF and the TAK/TAB complex interact, I once again included the relevant in silico interaction in Model A and performed WT and MyD88 KO simulations to LPS stimulus. In order to generate the delayed activation of MAP kinases observed in MyD88 KO conditions, I added at least two hypothetical intermediates between TRIF and the TAK/ TAB complex (Fig. 2A, Model B, Appendix A) in the same manner I performed earlier.6 Using the modified model I simulated in silico TAK1 KO for JNK activation in TLR4 activation. The model simulation showed complete abolition of JNK activation (Fig. 1B). Similar result for TAK1 KO cells was reported by Shim et al, 2005. Next, I performed in silico TRAF-6 KO simulations and reproduced p38 and NF-κB activity profile in accordance with Gohda et al, 2004 (Fig. 2C,D).

Figure 1B-C. A) (viewed on previous page) Reference Model TLR4 Model.

Figure 1B-C

A) (viewed on previous page) Reference Model TLR4 Model. The MyD88-dependent and MyD88-independent signaling pathways with hypothetical intermediates, adapted from Selvarajoo, 2006 B) The addition of a crosstalk mechanism between TRIF and TRAF6 molecules (Model (more...)

A computational model is acceptable only if it is able to predict multiple perturbation studies. The updated Model B is also able to predict the NF-κB activation for various types of available KO studies (MyD88 KO, TRAF-6-KO and TRIF KO) in macrophages (Fig. 2D).5,9,14 In all conditions and also for JNK and p38 relative activity, I observed the simulation profiles yield consistent result with experimental observations (Fig. 2B-D).


Recent studies have shown that the MyD88-dependent and -independent signaling cascades may interact and thereby possibly coregulate NF-κB and MAP kinases. For instance, blocking the MyD88 pathway through MyD88 KO in TLR4 stimulus also leads to JNK and NF-κB activation, albeit with delayed kinetics.4,9 There have been other reports that TRAF-6 binds to TRIF and may thus activate NF-κB in a MyD88-independent manner.3 However, cells deficient of TRAF-6 still showed JNK and NF-κB activation.4,5

Previously I reported a computational model (MyD88-dependent and -independent pathways) that demonstrated the probable reasons for the delayed NF-κB activation observed in MyD88-deficient mice.6 I used this model, with appropriate modifications, to first test the idea whether TRAF-6 binding to TRIF activates NF-κB (Model A, Fig. 1B). By performing in silico simulations and comparing the results with reported experimental findings, I showed that this mechanism is possible. However it may not be sufficient to concur that this is the only other MyD88-independent activator of NF-κB on top of TRIF to cRel via TBK1 mentioned in my previous model (Fig. 1A).

Figure 2C-D. A) Model B.

Figure 2C-D

A) Model B. The MyD88-dependent and MyD88-independent signaling pathways with the crosstalk mechanism between TRIF and TAK/TAB complex via a few hypothetical intermediates (protein, protein complexes or phosphorylation step). See Appendix A for details. (more...)

I next show that interactions possibly exist between the MyD88-dependent and -independent pathways downstream of TRAF-6 and upstream of MAP kinases. The Model B (which include the interaction between (i)TRIF and the TAK/TAB complex with numerous unclassified intermediates (protein, protein complexes or phosphorylation state) and the (ii)TRIF/ TBK1 pathway) recapitulated NF-κB and JNK activities for several KO conditions with reasonable simulations (Fig. 1B-D). In summary, my model suggests that there is possibly one additional crosstalk pathway, from TRIF to the TAK/TAB complex, for the activation of NF-κB and MAP kinases. However, further experimental studies are required to determine the intermediates that participate through the two suggested pathways.

It is known that TBK1 activates IFN-inducible genes via TRIF-dependent signaling.3 TBK1 associate with TANK15 and phosphorylate IRF3 in response to viral infection.16,17 TRAF3, which has a similar structure to TRAF6, has been recently reported to bind with TRIF.18 RIP1 has been implicated in activating the IKK complex19 and shown to mediate the recruitment of TAK1 to the TNF-R1 complex.20 Also, RIP1 has been shown to mediate TRIF-dependent, TLR4-induced NF-κB but not IRF-3 activation.21 Collectively, it appears that both TRAF3 and RIP1 might participate through the TRIF to TAK/TAB pathway (Fig. 3).

Figure 2A-B. See following page for figure legend.

Figure 2A-B

See following page for figure legend.


The TLR field is rapidly evolving and many time-course experiments are being performed to understand the regulatory roles of various signaling adaptors and molecules. Without the use of appropriate analytical tools, like the computational TLR4 pathway models, it is a daunting challenge for biologists to analyse and interpret the complex information generated by the various experiments. In this paper, I have shown the utility of in silico models to put together and test various hypotheses regarding the TLR signaling mechanism. I investigated, through various simulations, the crosstalk mechanisms between the MyD88-dependent and -independent pathways at early time signaling to LPS stimulus, without assuming too many details of the TLR4 signaling pathways from literature. The final Model B is consistent across several literature observations, showing how computer models can help us understand the mechanistic process behind the complex signaling dynamics. Such result from systemic work will surely provide hints to the wet-bench experimentalist to perform more targeted research that will eventually, but at an increasing pace, lead to the discovery of novel intracellular targets, say, for the TLR signaling in disease conditions.

Figure 3. The new proposed TLR4 pathway.

Figure 3

The new proposed TLR4 pathway. The MyD88-dependent and MyD88-independent signaling pathways with various investigated crosstalk mechanisms.

Appendix A

Table 1. Model A.

Table 1

Model A.

Table 2. Model B (The modifications/additions made for Model A).

Table 2

Model B (The modifications/additions made for Model A).


Koichi Matsuo, Masa Tsuchiya (Keio University) and Shizuo Akira (Osaka University) are thanked for their critical reading of the manuscript. I would also like to thank my funding agency, Biomedical Research Council of Singapore.


O'Neill LA. TLRs: Professor Mechnikov, sit on your hat. Trends Immunol. 2004;25:687–693. [PubMed: 15530840]
Akira S, Takeda K. Toll-like receptor signalling. Nat Rev Immunol. 4(7):499–511. [PubMed: 15229469]
Sato S, Sugiyama M, Yamamoto M. et al. Toll/IL-1 receptor domain-containing adaptor inducing IFN-beta (TRIF) associates with TNF receptor-associated factor 6 and TANK-binding kinase 1 and activates two distinct transcription factors, NF-kappa B and IFN-regulatory factor-3, in the Toll-like receptor signaling. J Immunol. 2003;171(8):4304–4310. [PubMed: 14530355]
Kawai T, Takeuchi O, Fujita T. et al. Lipopolysaccharide Stimulates the MyD88-Independent Pathway and Results in Activation of IFN-Regulatory Factor 3 and the Expression of a Subset of Lipopolysaccharide-Inducible Genes. J Immunol. 2001;167(10):5887–5894. [PubMed: 11698465]
Gohda J, Matsumura T, Inoue J. Cutting edge: TNFR-associated factor (TRAF) 6 is essential for MyD88-dependent pathway but not toll/IL-1 receptor domain-containing adaptor-inducing IFN-beta (TRIF)-dependent pathway in TLR signaling. J Immunol. 2004;73(5):2913–2917. [PubMed: 15322147]
Selvarajoo K. Discovering differential activation machinery of the Toll-like receptor 4 signaling pathways in MyD88 knockouts. FEBS Lett. 2006;580(5):1457–1464. [PubMed: 16458306]
Covert MW, Leung TH, Gaston JE. et al. Achieving stability of lipopolysaccharide-induced NF-kappaB activation. Science. 2005;309(5742):1854–1857. [PubMed: 16166516]
Yamada S, Shiono S, Joo A. et al. Control mechanism of JAK/STAT signal transduction pathway. FEBS Lett. 2003;534(1-3):190–196. [PubMed: 12527385]
Kawai T, Adachi O, Ogawa T. et al. Unresponsiveness of MyD88-deficient mice to endotoxin. Immunity. 1999;11(1):115–122. [PubMed: 10435584]
Hoffmann A, Levchenko A, Scott ML. et al. The IκB-NF-κB Signaling Module: Temporal Control and Selective Gene Activation. Science. 2002;298(5596):1241–1245. [PubMed: 12424381]
Takahashi K, Ishikawa N, Sadamoto Y. et al. E-Cell 2: multi-platform E-Cell simulation system. Bioinformatics. 2003;19:1727–1729. [PubMed: 15593410]
Shim JH, Xiao C, Paschal AE. et al. TAK1, but not TAB1 or TAB2, plays an essential role in multiple signaling pathways in vivo. Genes Dev. 2005;19(22):2668–2681. [PMC free article: PMC1283960] [PubMed: 16260493]
Sato S, Sanjo H, Takeda K. et al. Essential function for the kinase TAK1 in innate and adaptive immune responses. Nat Immunol. 2005;6(11):1087–1095. [PubMed: 16186825]
Hoebe K, Du X, Georgel P. et al. Identification of Lps2 as a key transducer of MyD88-independent TIR signalling. Nature. 2003;424(6950):743–748. [PubMed: 12872135]
McWhirter SM, Fitzgerald KA, Rosains J. et al. IFN-regulatory factor 3-dependent gene expression is defective in Tbk1-deficient mouse embryonic fibroblasts. Proc Natl Acad Sci USA. 2004;101(1):233–238. [PMC free article: PMC314168] [PubMed: 14679297]
Fitzgerald KA, McWhirter SM, Faia KL. et al. IKKepsilon and TBK1 are essential components of the IRF3 signaling pathway. Nat Immunol. 2003;4(5):491–496. [PubMed: 12692549]
Sharma S, tenOever BR, Grandvaux N. et al. Triggering the interferon antiviral response through an IKK-related pathway. Science. 2003;300(5622):1148–1151. [PubMed: 12702806]
Hacker H, Redecke V, Blagoev B. et al. Specificity in Toll-like receptor signalling through distinct effector functions of TRAF3 and TRAF6. Nature. 2005;439(7073):204–207. [PubMed: 16306937]
Poyet JL, Srinivasula SM, Lin JH. et al. Activation of the Ikappa B kinases by RIP via IKKgamma /NEMO-mediated oligomerization. J Biol Chem. 2000;275(48):37966–37977. [PubMed: 10980203]
Blonska M, Shambharkar PB, Kobayashi M. et al. TAK1 is recruited to the tumor necrosis factor-alpha (TNF-alpha) receptor 1 complex in a receptor-interacting protein (RIP)-dependent manner and cooperates with MEKK3 leading to NF-kappaB activation. J Biol Chem. 2005;280(52):43056–43063. [PubMed: 16260783]

Guest editor: Sankar Ghosh

Copyright © 2000-2013, Landes Bioscience.
Bookshelf ID: NBK6625


  • PubReader
  • Print View
  • Cite this Page

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...