Lipopolysaccharides from atherosclerosis-associated bacteria antagonize TLR4, induce formation of TLR2/1/CD36 complexes in lipid rafts and trigger TLR2-induced inflammatory responses in human vascular endothelial cells

Authors

  • Martha Triantafilou,

    1. Infection and Immunity Group, School of Life Sciences, University of Sussex, Falmer, Brighton, BN1 9QG, UK.
    Search for more papers by this author
  • Frederick G. J. Gamper,

    1. Infection and Immunity Group, School of Life Sciences, University of Sussex, Falmer, Brighton, BN1 9QG, UK.
    Search for more papers by this author
  • Philipp M. Lepper,

    1. Department of Intensive Care Medicine, University of Bern, Inselspital, Bern, Switzerland.
    2. Department of Internal Medicine II, University of Ulm, 89075 Ulm, Germany.
    Search for more papers by this author
  • Marios Angelos Mouratis,

    1. Infection and Immunity Group, School of Life Sciences, University of Sussex, Falmer, Brighton, BN1 9QG, UK.
    Search for more papers by this author
  • Christian Schumann,

    1. Department of Internal Medicine II, University of Ulm, 89075 Ulm, Germany.
    Search for more papers by this author
  • Evlambia Harokopakis,

    1. Center for Oral Health and Systemic Disease and Departments of Periodontics and Microbiology/Immunology and of Orthodontics/Pedodontics, University of Louisville Health Sciences Center, Louisville, KY 40292, USA.
    Search for more papers by this author
  • Robert E. Schifferle,

    1. Department of Oral Biology, University at Buffalo, NY 14214, USA.
    Search for more papers by this author
  • George Hajishengallis,

    1. Center for Oral Health and Systemic Disease and Departments of Periodontics and Microbiology/Immunology and of Orthodontics/Pedodontics, University of Louisville Health Sciences Center, Louisville, KY 40292, USA.
    Search for more papers by this author
  • Kathy Triantafilou

    Corresponding author
    1. Infection and Immunity Group, School of Life Sciences, University of Sussex, Falmer, Brighton, BN1 9QG, UK.
      *E-mail K.Triantafilou@sussex.ac.uk; Tel. (+44) 1273 678362; Fax (+44) 1273 678362.
    Search for more papers by this author

*E-mail K.Triantafilou@sussex.ac.uk; Tel. (+44) 1273 678362; Fax (+44) 1273 678362.

Summary

Infection with bacteria such as Chlamydia pneumonia, Helicobacter pylori or Porphyromonas gingivalis may be triggering the secretion of inflammatory cytokines that leads to atherogenesis. The mechanisms by which the innate immune recognition of these pathogens could lead to atherosclerosis remain unclear. In this study, using human vascular endothelial cells or HEK-293 cells engineered to express pattern-recognition receptors (PRRs), we set out to determine Toll-like receptors (TLRs) and functionally associated PRRs involved in the innate recognition of and response to lipopolysaccharide (LPS) from H. pylori or P. gingivalis. Using siRNA interference or recombinant expression of cooperating PRRs, we show that H. pylori and P. gingivalis LPS-induced cell activation is mediated through TLR2. Human vascular endothelial cell activation was found to be lipid raft-dependent and to require the formation of heterotypic receptor complexes comprising of TLR2, TLR1, CD36 and CD11b/CD18. In addition, we report that LPS from these bacterial strains are able to antagonize TLR4. This antagonistic activity of H. pylori or P. gingivalis LPS, as well as their TLR2 activation capability may be associated with their ability to contribute to atherosclerosis.

Introduction

Atherosclerosis is a chronic inflammatory disease that is modulated by both genetic and environmental factors (Libby, 2002; Mullaly and Kubes, 2004). Although it was formerly considered a lipid storage disease, nowadays it is believed to be triggered by several factors, which include hypertension, high plasma concentrations of low-density lipoprotein (LDL), cholesterol, diabetes and even infection (Ross, 1999).

Damage to the vascular endothelium is crucial in the development and progression of the disease. Common risk factors include hypercholesterolemia, hypertension, obesity and smoking. Although hypercholesterolemia seems to be a prerequisite for atherosclerogenesis most individuals with proven coronary artery disease have been found to have ‘average’ levels of cholesterol (Libby et al., 2002). In addition, even in cases of extreme hypercholesterolemia there is diversity in the expression of the disease (Leitersdorf et al., 1990), leading us to question what accounts for the differences in the rate of lesion formation and clinical presentation of atherosclerosis. It is becoming increasingly apparent that the aetiology is multifactorial, and that in addition to the ‘traditional’ risk factors (such as hypercholesterolemia, hypertension and smoking), we have newly emerging ‘non-traditional risk factors’, such as chronic infection and the potential role that the immune system might play in atherogenesis (Hansson, 2001; Hansson et al., 2002).

The first step in atherogenesis is the infiltration and entrapment of LDL in the blood vessel wall. LDL oxidizes readily and storage of oxidized LDL (oxLDL) represents the first main phase in the atherosclerotic mechanism. Subsequently, monocytes and macrophages are recruited to the lesions and this is typically an inflammatory event associated with increased cytokine levels. Endothelial cells, macrophages, T-cells and smooth muscle cells are the main cell types that are found in atherosclerotic lesions.

If we look at the mechanism of atherosclerosis, we must consider LDL receptors, called ‘scavenger A and B′ as the key molecules for the whole process. However, recently it has been suggested that infectious agents may initiate or promote the inflammatory process in atherosclerosis. In particular, it is believed that the lipopolysaccharide (LPS) from bacteria such as Chlamydia pneumonia (Kuo et al., 1995; Grayston, 2000; Schumacher et al., 2002), Helicobacter pylori (Danesh et al., 1997; 1999) or Porphyromonas gingivalis (Hajishengallis et al., 2002a; Gibson et al., 2006; Roth et al., 2006) may be triggering the secretion of inflammatory cytokines that leads to the recruitment of monocytes/macrophages to the lesions. Elevated levels of LPS seem to present a risk factor for the development of atherosclerosis (Kiechl et al., 2001), whereas injection of LPS has been shown to accelerate formation of atherosclerotic lesions (Lehr et al., 2001). Thus, in addition to LDL receptors, we have to start considering receptors involved in bacterial recognition as key molecules for the generation or acceleration of atherosclerosis.

Toll-like receptor 4 (TLR4), a key molecule in LPS recognition and signalling, has been shown to play a role in atherosclerosis (Zeuke et al., 2002). TLRs have been found to be expressed in human atherosclerotic lesions (Frantz et al., 1999; Edfeldt et al., 2002), whereas TLR4 polymorphism that attenuates receptor signalling is associated with decreased atherosclerosis in humans (Kiechl et al., 2002). The recent finding of TLR1, TLR2, TLR4 and TLR5 expression in both human and murine models of atherosclerosis (Edfeldt et al., 2002) offers a possible mechanism by which bacterial products might trigger atherogenesis. In addition, a polymorphism in CD14, the non-transmembrane receptor for LPS, that results in significantly higher density of CD14 on monocytes has been identified as a risk factor for myocardial infarction (Hubacek et al., 1999), suggesting that innate responses to infectious agents may participate in the initiation of atherosclerosis. There seems to be ample evidence that antigens derived from infectious pathogens, such as bacteria and viruses, could be triggering the innate immune system through pattern-recognition receptors (PRRs) (such as CD14 and TLRs) leading to immune activation in the atherosclerotic lesion.

Although TLRs play a crucial role in triggering innate and inflammatory responses, cellular activation by microbial molecules almost invariably involves interactions with several co-operating host receptors within membrane microdomains known as lipid rafts (Triantafilou et al., 2002). In the current study, we set out to determine the cooperating TLRs and PRRs involved in the innate recognition of LPS from bacterial pathogens, such as H. pylori and P. gingivalis, which are linked with atherosclerosis. Using HEK-293 cells transfected with specific TLRs, we demonstrate that H. pylori and P. gingivalis LPS-induced cell activation is mediated via TLR2. Cell activation by H. pylori and P. gingivalis LPS seem to require lipid raft function and formation of heterotypic receptor complexes comprising of TLR2, TLR1, CD36 and CD11b/CD18. In addition, our results suggest that LPS from these bacteria can act as antagonists for human TLR4. Parallel studies conducted with human vascular endothelial cells further confirm these findings. The unique ability of these bacterial strains to trigger signalling via TLR2, but antagonize TLR4, might shed more light into the reasons why these organisms develop chronic, low-level inflammation resulting in atherosclerosis.

Results

Helicobacter pylori and P. gingivalis LPS activation of human vascular endothelial cells

Cytokines have been proposed to play an important role in H. pylori- or P. gingivalis-induced atherosclerosis, but the exact mechanism of the cytokine induction remains unclear. Monocytes and macrophages are recruited to the atherosclerotic lesions and this is typically an inflammatory event that shows increased number of cytokines. Endothelial cells, macrophages, T-cells and smooth muscle cells are the main cell types that are found in lesions. Thus, we investigated whether H. pylori or P. gingivalis LPS can induce inflammatory cytokines in human vascular endothelial cells.

Human vascular endothelial cells were incubated with 1 μg ml−1H. pylori LPS. It was shown that H. pylori LPS could induce TNF-α and IL-6 cytokine production (P < 0.05; Fig. 1, black bar charts) in these cells, however, the cytokine production remained at low levels and was significantly lower than the cytokine release triggered by Escherichia coli LPS.

Figure 1.

Helicobacter pylori and P. gingivalis LPS-induced cytokine release in human vascular endothelial cells. Human vascular endothelial cells were incubated with no stimulus (white bar charts), 1 μg ml−1 of H. pylori LPS (black bar charts), 1 μg ml−1 of P. gingivalis LPS (grey bar charts) or 100 ng ml−1 of E. coli LPS (stripped bar charts). The supernatants were harvested and assayed for cytokine contents using the Cytometric Bead Array (CBA) system (Becton Dickinson). Fluorescence was detected using a FACSCalibur (BectonDickinson). The data presented are the mean of three independent experiments. Asterisks indicate statistical significance (P < 0.05).

Furthermore we proceeded to investigate whether P. gingivalis LPS can trigger cytokine production from vascular endothelial cells. We incubated human vascular endothelial cells with 1 μg ml−1 of P. gingivalis LPS. It was shown that P. gingivalis LPS could trigger TNF-α and IL-6 release from human vascular endothelial cells (P < 0.05; Fig. 1, grey bar charts), although the cytokine release was significantly weaker when compared with E. coli LPS (P < 0.05; Fig. 1, stripped bar charts).

Helicobacter pylori and P. gingivalis LPS-induced activation is TLR2-dependent

In order to investigate which TLRs might play a role in H. pylori or P. gingivalis LPS-induced activation of human vascular endothelial cells we used transfected cell lines. Human embryonic kidney (HEK) cells transfected with either TLR2, TLR3, TLR4/MD2 or TLR7 were used. HEK cells, which do not express TLRs, were found not to be able to produce TNF-α or IL-6 in response to H. pylori or P. gingivalis LPS (Fig. 2). Similarly, H. pylori or P. gingivalis LPS did not trigger cytokine production in HEK cells transfected with TLR3, TLR4 or TLR7 (Fig. 2). On the contrary, HEK cells transfected with TLR2 produced TNF-α and IL-6 after incubation with H. pylori or P. gingivalis LPS (P < 0.05; Fig. 2). Control cultures were stimulated with established TLR2 (Staphylococcus aureus LTA) or TLR4 (E. coli LPS) ligands and results were as expected (P < 0.05; Fig. 2).

Figure 2.

TLR2-dependent cytokine secretion in response to H. pylori and P. gingivalis LPS. HEK-293 cells transfected with TLR2, TLR3, TLR4 or TLR7 were incubated with no stimulus (white bar charts), 1 μg ml−1 of H. pylori LPS (black bar charts), 1 μg ml−1 of P. gingivalis LPS (grey bar charts), 100 ng ml−1 of E. coli LPS (stripped bar charts) or 10 μg ml−1S. aureus LTA (chequered bar charts). The supernatants were harvested and assayed for cytokine content using the Cytometric Bead Array (CBA) system (Becton Dickinson). Fluorescence was detected using a FACSCalibur (BectonDickinson). The data represent the mean of three independent experiments. Asterisks indicate statistical significance (P < 0.05).

NF-κB transcriptional responses to H. pylori and P. gingivalis LPS are mediated by TLR2

TLRs act upstream of NF-κB activation. TLR signalling pathways have been shown to ultimately result in the release of NF-κB from its endogenous inhibitor (Akira, 2001) and subsequent nuclear translocation that leads to the transcription of inflammatory cytokines.

In order to determine whether H. pylori or P. gingivalis LPS-induced TLR2 activation that we observed leads to NF-κB-driven transcription response, we used HEK cells transfected with either TLR2, TLR3, TLR4 or TLR7 and an NF-κB luciferase reporter gene. Cells were stimulated as indicated and after 6 h of stimulation, the cells were lysed in passive lysis buffer (Promega). Luciferase activity was measured using a plate reader luminometer. Similar to our previous findings, it was shown that H. pylori (black bar charts) or P. gingivalis (grey bar charts) LPS were not able to induce NF-κB activation in HEK-TLR3, TLR4 or TLR7 cells. In contrast, H. pylori or P. gingivalis LPS could induce NF-κB activation only in cells, which expressed TLR2 (P < 0.05; Fig. 3).

Figure 3.

TLR2-dependent NF-κB activation in response to H. pylori and P. gingivalis LPS. HEK-293 cells transfected with TLR2, TLR3, TLR4 or TLR7 were incubated with no stimulus (white bar charts), 1 μg ml−1 of H. pylori LPS (black bar charts), 1 μg ml−1 of P. gingivalis LPS (grey bar charts), 100 ng ml−1 of E. coli LPS (stripped bar charts) or 10 μg ml−1S. aureus LTA (chequered bar charts). After stimulation, the cells were lysed and analysed for luciferase activity. The data shown represent a mean of three independent experiments. Asterisks indicate statistical significance (P < 0.05).

Inhibition of H. pylori or P. gingivalis LPS-induced activation of human vascular endothelial cells by silencing TLR2

Because we had already demonstrated the importance of TLR2 in H. pylori or P. gingivalis LPS-mediated activation in transfected cell lines, we investigated whether TLR2 was also important for triggering the inflammatory response in human vascular endothelial cells.

In order to determine the role of TLR2 in H. pylori or P. gingivalis LPS recognition, we used RNA interference (siRNA) to knock down the expression of TLR2 in human vascular endothelial cells. Transfection with synthetic TLR2-specific psiRNA resulted in 60% decrease in TLR2 as determined by Western blotting (Fig. 4A). Control transfections of human vascular endothelial cells with the psiRNA vector did not affect TLR2 expression.

Figure 4.

Inhibition of H. pylori and P. gingivalis LPS-induced activation of human vascular endothelial cells by silencing TLR2.
A. RNA interference resulted in decreased expression of TLR2 and TLR7 as determined by Western blotting.
B. Following RNA interference, human vascular endothelial cells were not stimulated (white bar charts) or incubated with 1 μg ml−1 of H. pylori LPS (black bar charts), 1 μg ml−1 of P. gingivalis LPS (grey bar charts), 100 ng ml−1 of E. coli LPS (stripped bar charts) or 10 μg ml−1S. aureus LTA (chequered bar charts). The supernatants were harvested and assayed for cytokine secretion using the Cytometric Bead Array (CBA) system (Becton Dickinson). Fluorescence was detected using a FACSCalibur (BectonDickinson). The data represent the mean of three independent experiments. Asterisks indicate statistical significance (P < 0.05).

Following RNA interference, human vascular endothelial cells were incubated with, H. pylori (Fig. 4B, black bar charts) or P. gingivalis LPS (Fig. 4B, grey bar charts). Cytokine assays were performed after the designated incubation times. It was shown that silencing of TLR2 inhibited H. pylori or P. gingivalis LPS-induced cellular activation (Fig. 4B), thus suggesting the importance of TLR2 in H. pylori or P. gingivalis LPS-induced activation of human vascular endothelial cells. Control experiments were performed by employing psiRNA technology in order to silence TLR7. Transfection with synthetic TLR7-specific psiRNA resulted in 60% decrease in TLR7 expression as determined by luciferase activity and Western blotting. Following RNA interference for TLR7, human vascular endothelial cells were incubated with, H. pylori or P. gingivalis LPS. Silencing of TLR7 did not affect the production of TNF-α in response to H. pylori or P. gingivalis LPS, thus suggesting that the inhibition observed when TLR2 was silenced was specific.

Helicobacter pylori and P. gingivalis LPS-induced receptor clusters

Because it has recently been demonstrated that TLR4/MD-2 clustering is crucial for LPS-induced signalling (Visintin et al., 2003), and we have demonstrated that recruitment of TLR4/MD-2 within different combinational clusters determines immune response against different types of LPS (Triantafilou et al., 2004a), we proceeded to investigate whether H. pylori LPS and P. gingivalis LPS induce the formation of similar activation clusters on human vascular endothelial cells. We had previously investigated the formation of similar clusters in response to H. pylori LPS on human monocytes (Lepper et al., 2005).

We measured FRET in terms of dequenching of donor fluorescence after complete photobleaching of the acceptor fluorophore. We tested the energy transfer efficiency in our system using a positive control, i.e. energy transfer between mAbs to different epitopes on CD14 molecules, showing that the maximum energy transfer efficiency (E%) was 38 ± 1.0. A negative control between Cy3-CD14 and Cy5-W6/32 [mAb specific for major histocompatibility complex class I] was also used, which revealed no significant energy transfer (6 ± 1.0). We proceeded to measure FRET between TLR2 and different receptor molecules that have been implicated in TLR2-dependent activation (TLR1, TLR6, CD36, CD11b/CD18), thus we measured FRET between TLR2 and these molecules in response to the two types of LPS. TLR2 was found not to associate with these receptor molecules prior to H. pylori or P. gingivalis LPS stimulation (Fig. 5).

Figure 5.

TLR2 heterotypic associations in response to H. pylori and P. gingivalis LPS. Human vascular endothelial cells were stimulated with no stimulus (white bar charts), 1 μg ml−1 of H. pylori LPS (black bar charts), 1 μg ml−1 of P. gingivalis LPS (grey bar charts), 100 ng ml−1 of E. coli LPS (stripped bar charts) or 10 μg ml−1S. aureus LTA (chequered bar charts). Energy transfer between TLR2 (Cy3) and the different receptors was measured from the increase in donor (Cy3) fluorescence after acceptor (Cy5) photobleaching. The percentage of energy transfer and standard deviation was calculated from three independent experiments. Asterisks indicate statistical significance (P < 0.05).

Energy transfer between TLR2-Cy3 and the various Cy5-labelled molecules was measured. TLR2 was found to associate with TLR1, CD11b/CD18 and CD36 after H. pylori LPS stimulation, whereas the same activation cluster was formed in response to P. gingivalis LPS, thus suggesting that P. gingivalis LPS triggers the formation of an identical activation cluster to the one observed in response to H. pylori LPS.

Control experiments using the method described by Kenworthy and Edidin (1998) ruled out the possibility that the FRET observed was due to randomly distributed molecules (data not shown).

Recruitment of TLR2 in lipid rafts following H. pylori or P. gingivalis LPS stimulation

It has been previously shown that regions of the plasma membrane known as lipid rafts, or microdomains facilitate LPS-induced cell activation (Wang et al., 1996; Triantafilou et al., 2002). Recently we have demonstrated that TLR2 is also recruited within lipid rafts upon stimulation by its ligands (Triantafilou et al., 2004b). Because we demonstrated that H. pylori and P. gingivalis LPS-induced cellular activation is mediated through TLR2, we proceeded to determine whether TLR2 was recruited within lipid rafts upon stimulation by H. pylori and P. gingivalis LPS. FRET experiments between TLR2 and GM-1 ganglioside were performed before and after stimulation by H. pylori and P. gingivalis LPS. TLR2 molecules were labelled with Cy3-TL2.1 and GM-1 ganglioside, a raft-associated lipid, was labelled with Cy5-cholera toxin. It was shown that similarly to LTA, H. pylori and P. gingivalis LPS could recruit TLR2 in lipid rafts (Fig. 6).

Figure 6.

TLR and GM-1 ganglioside FRET measurements before and after H. pylori or P. gingivalis LPS stimulation. TLR and GM-1 ganglioside FRET measurements before and after H. pylori or P. gingivalis LPS stimulation of human vascular endothelial cells. Energy transfer between TLR2 (Cy3-TLR2) or TLR4 (Cy3-TLR4) and GM-1 ganglioside (Cy5-cholera-toxin) before (white bar charts) and after stimulation with 1 μg ml−1 of H. pylori LPS (black bar charts), 1 μg ml−1 of P. gingivalis LPS (grey bar charts), 100 ng ml−1 of E. coli LPS (stripped bar charts) or 10 μg ml−1S. aureus LTA (chequered bar charts). Energy transfer between TLR2 (Cy3) and the different receptors was measured from the increase in donor (Cy3) fluorescence after acceptor (Cy5) photobleaching. The percentage of energy transfer and standard deviation was calculated from three independent experiments. Asterisks indicate statistical significance (P < 0.05).

In addition, TLR4 has been shown to be recruited within these microdomains after enterobacterial LPS stimulation and this clustering is crucial for LPS-induced cytokine production (Triantafilou et al., 2002). In order to test whether stimulation by H. pylori and P. gingivalis LPS affects the recruitment of TLR4 within lipid raft we performed FRET experiments between TLR4 and GM1-ganglioside, a lipid-raft marker. TLR4 molecules were labelled with Cy3-HTA125, MD2 was also labelled with Cy3-MD-2 polyclonal antibody and GM-1 ganglioside, a raft-associated lipid, was labelled with Cy5-cholera toxin. FRET measurements confirmed the presence of TLR4/MD-2 in lipid rafts after E. coli LPS stimulation. Similarly to our previous study, FRET measurements further demonstrated that there was less energy transfer between GM-1 and TLR4/MD-2 after H. pylori and P. gingivalis LPS stimulation (Fig. 6), suggesting that there is less recruitment of TLR4/MD-2 molecules within lipid rafts.

Inhibitory effects of H. pylori or P. gingivalis LPS on LPS-induced secretion of TNF-a

It has been previously suggested that LPS from non-enterobacteria, such as Rhodobacter sphaeroides (Loppnow et al., 1993), Rhodobacter capsulatus (Rose et al., 1995), P. gingivalis (Hajishengallis et al., 2002b; Yoshimura et al., 2002; Coats et al., 2003) and Capnocytophaga ochracea (Yoshimura et al., 2002) could act as potent TLR4 antagonists. In addition, we have previously shown that H. pylori LPS could act as a TLR4 antagonist as well (Lepper et al., 2005), thus we proceeded to investigate whether this was the case in human vascular endothelial cells. We pretreated human vascular endothelial cells with 1 μg ml−1 of either H. pylori or P. gingivalis LPS for 1 h, and subsequently stimulated the cells with 100 ng ml−1E. coli LPS. The addition of either H. pylori or P. gingivalis LPS attenuated the TNF-α secretion by human vascular endothelial cells stimulated with E. coli LPS (P < 0.05; Fig. 7), thus demonstrating that H. pylori and P. gingivalis LPS can act as an LPS antagonist. The suppressive effect of the combined LPS concentrations was not a result of toxicity, as demonstrated by trypan blue exclusion (data not shown).

Figure 7.

Inhibitory effects of H. pylori or P. gingivalis LPS on E. coli LPS-induced cytokine secretion. Human vascular endothelial cells were pretreated with 1 μg ml−1 of H. pylori or P. gingivalis LPS before stimulation with E. coli LPS (100 ng ml−1). The supernatants were harvested and assayed for cytokine content using the Cytometric Bead Array (CBA) system (Becton Dickinson). Fluorescence was detected using a FACSCalibur (BectonDickinson). The data represent the mean of three independent experiments. Asterisks indicate statistical significance (P < 0.05).

Inhibitory effects of H. pylori and P. gingivalis LPS on activation of HEK-CD14-TLR4/MD2 cells

In order to further determine whether H. pylori and P. gingivalis LPS were a TLR4 antagonist, we used HEK cells transfected with TLR3, TLR4 or TLR7 and an NF-κB luciferase reporter gene. Cells were stimulated as indicated and after 6 h of stimulation, the cells were lysed in passive lysis buffer (Promega). Luciferase activity was measured using a plate reader luminometer. When HEK-CD14-TLR4 cells were preincubated with H. pylori or P. gingivalis LPS (1 μg ml−1), it resulted in an attenuated NF-κB luciferase activity in response to E. coli LPS (Fig. 8), thus suggesting that H. pylori or P. gingivalis LPS specifically antagonizes TLR4.

Figure 8.

Inhibitory effects of H. pylori or P. gingivalis LPS on activation of HEK/CD14/TLR4 cells. HEK-293 cells transfected with TLR3, TLR4 or TLR7 were either not preincubated (white bar charts) or preincubated with H. pylori (black bar charts) or P. gingivalis (grey bar charts) LPS (1 μg ml−1), prior to stimulation with 100 ng ml−1 of E. coli LPS (stripped bar charts). After stimulation, the cells were lysed and analysed for luciferase activity. Asterisks indicate statistical significance (P < 0.05).

Discussion

Atherosclerosis was formerly considered a bland lipid storage disease; nowadays it has been revealed that it involves an ongoing inflammatory response. Recent studies have suggested that infectious agents may initiate or promote the inflammatory process in atherosclerosis. In particular, it is believed that the LPS from bacteria such as C. pneumonia (Kuo et al., 1995; Grayston, 2000; Schumacher et al., 2002), H. pylori (Danesh et al., 1997; 1999; Mayr et al., 2000) or P. gingivalis (Hajishengallis et al., 2002a; Gibson et al., 2006) may be triggering the secretion of inflammatory cytokines that leads to the recruitment of monocytes/macrophages to the lesions. These particular microorganisms are common aetiological agents of chronic infection in humans, but how they achieve this remains unclear. TLR4 a key molecule in LPS recognition and signalling has been shown to play a role in atherosclerosis (Zeuke et al., 2002). TLRs have been found to be expressed in human atherosclerotic lesions (Frantz et al., 1999; Edfeldt et al., 2002), whereas TLR4 polymorphism that attenuates receptor signalling is associated with decreased atherosclerosis in humans (Kiechl et al., 2002). The function of other TLRs or other bacterial receptors has not yet been investigated. The possibility that in addition to TLR4, other TLR molecules might be involved, is strengthened by recent findings that demonstrate that there is reduced atherosclerosis in MyD88-null mice (Bjorkbacka et al., 2004; Michelsen et al., 2004). MyD88 is an adaptor molecule that transduces signalling events downstream of the TLRs. Furthermore, TLR2 and not TLR4 are implicated in innate immune responses against LPS from C. pneumonia (Kuo et al., 1995; Grayston, 2000) and H. pylori (Lepper et al., 2005). Thus, there is a strong possibility that TLR2 might also be involved in infection-induced atherosclerosis.

In this study we explored the role of TLR2 and TLR4 in H. pylori and P. gingivalis LPS recognition in human vascular endothelial cells. By using transfected cell lines, we demonstrated that LPS from both bacteria that we tested activates vascular endothelial cells via TLR2.

Interestingly, when we investigated whether H. pylori and P. gingivalis LPS could induce the formation of receptor clusters on vascular endothelial cells, it was shown that they could induce receptor clusters comprising of TLR2, TLR1, CD36 and CD11b/CD18. These clusters seem to form within lipid rafts. There is a possibility that other receptors might associate with the cluster. Because these pathogens trigger responses via TLR2, it is possible that CD36 could act as a key molecule within the receptor cluster. CD36 has recently been shown to associate with TLR2 (Hoebe et al., 2005). Thus, in addition to bacterial products, TLR2 might be engaged by endogenous agonists, such as lipid-derived CD36 ligands, because they are in close proximity, which exacerbate the inflammatory response.

Because several non-enterobacterial LPS that activate cells via TLR2, have been shown to act as TLR4 antagonists, we proceeded to investigate whether H. pylori and P. gingivalis LPS could also act as TLR4 antagonists on vascular endothelial cells. Our FRET data suggested that P. gingivalis LPS may be a weak TLR4 agonist; this is consistent with the notion that P. gingivalis expresses a heterogeneous mixture of LPS resulting in TLR2 agonistic, weak TLR4 agonistic and strong TLR4 antagonistic activities (Dixon and Darveau, 2005). Similarly, H. pylori LPS was found to also act as a TLR4 antagonist.

Our findings that H. pylori and P. gingivalis LPS induce responses via TLR2, and not TLR4, and the fact that they are able to act as a TLR4 antagonists shed more light into the reasons why these organism can develop chronic, low-level inflammation of vascular endothelial cells leading to atherogenesis. Both types of LPS seem to be able to prevent a rigorous TLR4 response that would eliminate the bacteria, whereas they seem to stimulate a low-level continuous TLR2 activation that contributes to local chronic endothelial inflammation. There seems to be an interplay of several factors that aid H. pylori and P. gingivalis to induce this low level inflammation in vascular endothelial cells. In addition to the fact that H. pylori LPS may mimic Lewis blood group antigens, and thus ‘camouflage’ the organism, minimizing host defence, one of the reasons for this chronic, low level inflammation might be because it activates via TLR2. On the other hand, P. gingivalis possesses adhesive filamentous appendages on its cell surface, known as fimbriae. Fimbriae constitute a major P. gingivalis virulence factor, not only in terms of the colonization potential of the pathogen (Malek et al., 1994) but also for downregulating Il−12 via interactions with CD11b/CD18 (Hajishengallis et al., 2005). Similarly to H. pylori, P. gingivalis LPS is also sensed via TLR2. Activation via TLR2 results in a subdued inflammatory response (Hirschfeld et al., 2001) allowing the organism to establish a chronic foothold in the vascular endothelium. Furthermore, the TLR4 antagonistic activity that these bacteria display seems to provide them with even greater advantage over the host in order to escape from the host innate immune system. The antagonistic activity could possibly be a virulence factor for certain strains, and might be associated with the progression of H. pylori or P. gingivalis-induced atherosclerosis. Because TLR2 seems to be the central receptor involved in this activation, progression to atherosclerotic disease could be a combination of low-level chronic inflammatory responses triggered by microbial infections via TLR2, as well as engagement of TLR2 by CD36 lipid derived endogenous ligands, which must augment the response and result in the formation of the lesions.

Experimental procedures

Materials

All fine chemicals were obtained from Sigma (UK). TLR4-specific mAb, HTA125 and TLR2-specific mAb, TL2.1 were obtained from Hycult (UK). Cholera toxin was purchased from List Laboratories The antibodies used for FRET studies were conjugated to either Cy3 or Cy5 using the labelling kits from AmershamPharmacia (UK).

Cell lines

Human vascular endothelial cells were obtained from TCS Cellworks (UK). HEK-293, HEK-239 cells transfected with CD14 and TLR4/MD2 or CD14 and TLR2 were maintained in DMEM containing 4.6 g l−1 glucose with 10% FCS, 5 μg ml−1 puromycin and 0.5 mg ml−1 G418-sulphate.

Helicobacter pylori LPS extraction

Lipopolysaccharide was extracted using the hot-phenol-water method as described previously (Westphal and Jann, 1965). Briefly, bacteria were harvested in acetone, filtered and dried. The dry bacteria were resuspended in 20 ml of 68°C water. Twenty millilitres of phenol at 68°C were added and the suspension was stirred for 20 min. Afterwards the suspension was cooled to 10°C. The watery phase was separated and dialysed. Ethanol precipitation was performed and ultracentrifugation of supernatants lead to LPS, which was digested by proteinase K and lyophilized. In order to verify that LPS was free from contaminating lipoproteins, we purified LPS as previously described (Manthey and Vogel, 1994; Hirschfeld et al., 2000).

Porphyromonas gingivalis LPS extraction

Lipopolysaccharide from P. gingivalis strain 381 was highly purified by phenol-water extraction and subsequent treatment with DNase I, RNase A and proteinase K, followed by chromatographic purification using a column of Sephacryl S-400 HR (2.5 by 40 cm; Pharmacia Fine Chemicals, Piscataway, NJ) (Hajishengallis et al., 2002b). The purity of the preparation was confirmed by immunodiffusion analysis and SDS-PAGE with silver staining. Moreover, colloidal gold staining revealed no detectable protein contamination in any of the preparations (5 μg of purified LPS applied per lane).

Cytokine assays

Human vascular endothelial cells were stimulated with no stimulus or H. pylori LPS, P. gingivalis LPS, E. coli LPS or S. aureus LTA (kindly provided by Professor Thomas Hartung). The cultures were incubated for the designated times. The supernatants were collected and frozen until the cytokine assays were performed. The BectonDickinson bead array system was used in order to determine the level of multiple cytokines at the same time.

Luciferase reporter assays for NF-κB activation

Human embryonic kidney 293 cells transfected with either TLR2, TLR3, TLR4 or TLR7 were seeded into 96 well plates. The following day, the cells were transiently transfected with a NF-κB luciferase reporter gene using lipofectamine 2000 (Invitrogen, UK) according to the manufacturer's instructions. The next day the cells were stimulated as indicated and after 6 h of stimulation, the cells were lysed in passive lysis buffer (Promega). Luciferase activity was measured using a plate reader luminometer.

Cell labelling for FRET

Human vascular endothelial cells were labelled with 100 μl of a mixture of donor conjugated antibody Cy3 and acceptor conjugated antibody Cy5. The cells were either not stimulated, or stimulated with H. pylori LPS or P. gingivalis LPS, and were rinsed twice in PBS/0.02% BSA, prior to fixation with 4% formaldehyde for 15 min. The cells were fixed in order to prevent potential reorganization of the proteins during the course of the experiment.

Cells were imaged on a Carl Zeiss, LSM510 META confocal microscope (with an Axiovert 200 fluorescent microscope) using a 1.4 NA 63× Zeiss objective. The images were analysed using LSM 2.5 image analysis software (Carl Zeiss). Cy3 and Cy5 were detected using the appropriate filter sets. Using typical exposure times for image acquisition (less than 5 s), no fluorescence was observed from a Cy3-labelled specimen using the Cy5 filters, nor was Cy5 fluorescence detected using the Cy3 filter sets.

FRET measurements

FRET is a non-invasive imaging technique that can be used in order to study molecular associations. It involves non-radiative transfer of energy from the excited state of a donor molecule to an appropriate acceptor. The rate of energy transfer is inversely proportional to the sixth power of the distance, between donor and acceptor. The efficiency of energy transfer (E) is defined with respect to r and R0, the characteristic Forster distance by:

image(1)

In the present study, FRET was measured using a method as previously described (Bastiaens et al., 1996; Kenworthy and Edidin, 1998; Triantafilou et al., 2001).

Statistical analysis

Statistical analysis was performed using the program SigmaStat 2.03 (SPSS). Results were considered statistically significant if P-value was < 0.05.

Acknowledgements

This work was supported by Heart Research UK (to K.T. and M.T.) and by US Public Health Service Grant DE015254 (to G.H.).

Ancillary