Unit: Cytokines and inflammation
Director: Cavaillon Jean-Marc
We analyze the early phase of innate immunity, namely, the mechanisms involved in cytokine induction by Gram negative or Gram positive bacteria and their pathogen-associated molecular patterns (PAMPs). While this response is a prerequisite to fight infection, its overwhelming activation may be associated with deleterious effects as those seen during septic shock or non-infectious systemic inflammatory response syndrome (SIRS). On another hand, the counteracting anti-inflammatory response may lead to an altered immune status (figure 1). We study the production of cytokines and their involvement in different inflammatory disorders in humans (sepsis, trauma, ischemia/reperfusion, cystic fibrosis). We investigate at the intracellular signaling pathways level, the reprogrammation of circulatory cells, reminiscent of the endotoxin tolerance phenomenon, that could be linked to the increased susceptibility of these patients to nosocomial infections.
We carry out both fundamental researches and clinical investigations, the later being pursued thanks to the collaboration with doctors in different hospitals.
1- MOLECULAR BASIS OF LEUKOCYTE REPROGRAMMING DURING SYSTEMIC INFLAMMATION (M. Adib-Conquy)
Since we first reported the in vitro hyporeactivity of circulating monocytes in sepsis patients in terms of cytokine production (Muñoz et al. J. Clin. Invest. 1991, 88, 1747), we have further characterized the immune-depression associated to this pathology. We have described a similar alteration in patients after surgery (Cabié et al. Cytokine 1992, 4, 576) and in patients resuscitated after cardiac arrest (Adrie et al. Circulation, 2002, 106, 562). We extended our observation to circulating neutrophils (Marie et al. Blood 1998, 91, 3439) and T-lymphocytes (Muret et al. Shock 2000, 13, 169). We study the intracellular molecular mechanisms responsible for the immune-depression observed in sepsis patients as well as in patients with non-infectious SIRS. A global decrease of nuclear factor-κB (NF-κB), an unbalance between its active (p65p50) and inactive (p50p50) forms and a weak cytoplasmic expression of its inhibitor (IκBα) have been observed within mononuclear cells of sepsis patients (Adib-Conquy et al. Am. J. Respir. Crit Care Med. 2000, 162, 1877). A similar study undertaken in trauma patients revealed that the defect in NF-κB expression lasts for more than 10 days (Adib-Conquy et al. J. Leuk. Biol. 2001, 70, 30).
The immune dysregulation is illustrated by a dramatic decrease in the capacity of peripheral blood mononuclear cells (PBMC) to release tumor necrosis factor (TNF) upon activation by Escherichia coli endotoxin (lipopolysaccharide, LPS) and CpG oligonucleotide, and to produce interleukin-6 (IL-6) in response to IL-1 and TNF. However, this is not a global defect, since cells from non-infectious SIRS remain fully reactive to other stimuli (Staphylococcus aureus, Streptococcus pyogenes, and Toll-like receptor-2 (TLR2) agonist). We have shown that the surface expression of TLR2 was not reduced on patients monocytes as compared to healthy controls, whereas that of TLR4 was reduced. However, the hyporeactivity to Gram-negative bacteria and E. coli LPS cannot be fully explained by the down-regulation of TLR4. Indeed, PBMC produce enhanced levels of IL-10 and IL-1 receptor antagonist (IL-1ra) in response to E. coli LPS as compared to cells from healthy controls. The activation of the p38 MAPK and the Sp-1 transcription factor was increased in PBMC from trauma patients after E. coli LPS or heat-killed Staphylococcus aureus (HKSA), stimulation, and the addition of an inhibitor of p38 decreased IL-10 production. Furthermore, heterotrimeric Gi proteins and phosphatidylinositol-3'-kinase were involved in IL-10 production. In conclusion, the immune-dysregulation described in SIRS patients is not a generalized phenomenon, but depends on the stimuli and the signaling pathways (Adib-Conquy et al. Am. J. Respir. Crit Care Med. 2003, 168, 158).
More recently, we showed that the TNF production in response to Pam3CysSK4, a specific TLR2 ligand, was also reduced in sepsis as compared to healthy controls. However, responsiveness to HKSA, heat-killed Escherichia coli, and muramyl dipeptide (MDP) was similar to that obtained with cells from healthy donors. MDP is sensed by Nod2 intracellular receptor. We found that the expression of Nod2 mRNA was unchanged in sepsis, as well as that of Nod1, a sensor of peptidoglycan from Gram-negative bacteria. In contrast to TNF, the production of IL-10 by monocytes of septic patients in response to LPS and Pam3CysSK4 was enhanced, whereas that induced by heat-killed bacteria was unchanged. We investigated the expression of various molecules regulating negatively the TLR4-dependent signaling. We showed by RT-PCR that in monocytes of sepsis patients, the expression of mRNA for Tollip and SOCS1 was similar to that of healthy controls, while MyD88s and SIGIRR expression was significantly enhanced (figure 2). We also included a group of resuscitated patients after cardiac arrest (RCA), in order to compared sepsis with non-infectious SIRS. Their monocytes responsiveness displayed a reduced production of TNF only in response to LPS, but not Pam3CysSK4 nor heat-killed bacteria. In contrast to sepsis patients, there was no increase in IL-10 production after ex vivo stimulation. The expression of SIGIRR mRNA was increased in the monocytes from RCA patients, but not that of MyD88s. Thus, leukocytes hyporesponsiveness to microbial agonists is not a generalized phenomenon in sepsis and SIRS, and the terms "leukocyte reprogramming" better define this phenomenon, rather than the terms often employed such as anergy, immunodepression, or immunoparalysis. This work is carried out with the collaboration of Dr. Christophe Adrie (Hôpital Delafontaine, Saint-Denis) and Rudi Beyaert (Ghent University, Belgium).
2- PRODUCTION OF MIF (V. Maxime, C. Fitting)
Macrophage migration inhibitory factor (MIF) is a pro-inflammatory cytokine that prevents glucocorticoid activities. In collaboration with Prof., D. Annane (Hôp. Poincaré, Garches) we have studied the capacity of leukocytes of sepsis patients versus those of healthy controls to produce MIF in vitro in response to various activators, and assessed the effect of a glucocorticoid treatment of the patients on the in vitro production of MIF by their leukocytes. PBMC from patients contained significantly higher amounts of MIF than cells from healthy controls. In culture, spontaneous release and release induced by LPS, HKSA and red blood cell lysates were significantly higher in patients than in controls. PBMC from patients treated with glucocorticoids showed a lower release of MIF in response to LPS, heat killed Escherichia coli and peptidoglycan than PBMC from untreated patients, and had levels similar to those obtained with PBMC from healthy controls. Thus, MIF is the first pro-inflammatory cytokine, of which ex vivo release by circulating cells is shown to be enhanced in sepsis. In contrast to the knowledge derived from animal studies, glucocorticoid treatment led to normalize the release of MIF by circulating PBMC from patients with septic shock (V. Maxime et al. J. Infect. Dis. 2005, 191, 138).
3- CYTOKINES AND CYSTIC FIBROSIS (A-F. Petit-Bertron, M. Adib-Conquy, C. Fitting)
We study neutrophils (PMN) derived from sputum of young patients with cystic fibrosis (CF). CF PMN display a high ex vivo spontaneous IL-8 production which can neither be up-regulated by the addition of LPS nor down-regulated by dexamethasone in contrast to what was observed in blood neutrophils (Corvol et al. Am. J. Physiol. 2003, 284, L997). As well, IL-10, that inhibits IL-8 production by blood PMN activated by LPS or peptidoglycan, has no effect on sputum PMN. In CF patients, TLR2 expression is significantly reduced on blood PMN from CF patients and TLR4 expression is significantly enhanced on sputum PMN. The expression of CD64 (Fcγ receptor type I), a marker of cell activation, is increased on blood PMN of CF patients. PMN from sputum of CF patients, kept in contact with respiratory epithelial cells, display an enhanced survival (reduced apoptosis) as compared to blood PMN from patients and healthy controls. A synergy in terms of IL-6 and IL-8 production is observed when PMN from CF patients are cultured in the presence of CF bronchial epithelial cells as compared to normal bronchial epithelial cells. This work is performed in collaboration with Drs J. Jacquot, O. Tabary, H. Corvol and A. Clément (Hôp. Trousseau, Paris) and founded by Association Vaincre la Mucovosicidose.
4- CELL SIGNALING IN MACROPHAGES VIA TOLL-LIKE RECEPTORS-2 & 4 AND NOD1 (R. Kapetanovic, C. Fitting, M. Adib-Conquy)
In collaboration with Dana Philpott (Immunité Innée & Signalisation, Inst. Pasteur), we investigate the role of TLR2, TLR4, and Nod1 & 2 molecules in cellular signaling which specifically leads to the production of TNF and IL-10.
Toll-like receptors (TLR) are molecules with highly conserved structures involved in the sensing of microbial-derived compounds. Nod1 and Nod2 are intracytoplasmic receptors and recognize distinct peptidoglycan-derived muropeptides. Macrophages possess phagocytic properties and their response to whole bacteria may be initiated both by TLR at the cell surface, and intracellularly by Nod proteins. We analyzed the relative contribution of these receptors to pro- and anti-inflammatory cytokines production by macrophages, following stimulation with whole bacteria. Using knockout mice, we showed that TLR4 and TLR2 contribution was predominant in the induction of TNF and IL-10 by Gram-negative bacteria. In contrast, the absence of TLR2 (and/or TLR4) did not affect the response to Gram-positive bacteria. However, when lacking TLR2, the phagocytic property of macrophages was essential to cytokine production in response to HKSA, that was markedly inhibited by cytochalasin D. By transfecting Raw 264.7 cells with dominant negative forms of Nod1 and Nod2, we showed that both molecules inhibited NF-κ B activation in response to HKSA. The interference of Nod1 in the response to Nod2 agonists was also confirmed by overexpression experiments in HEK293T cells, and may be explained by the capacity of Nod1 and Nod2 to form heterodimers, as shown by immunoprecipitation experiments. Finally, LPS and HKSA upregulate the expression of Nod1 and/or Nod2 mRNA, demonstrating a cross-talk at the transcriptional level between Nod1 and Nod2 in one hand, and between these intracytoplasmic receptors and TLRs in the other hand.
5- ENDOTOXIN TOLERANCE (C. Fitting, M. Adib-Conquy)
Endotoxin tolerance is characterized by a decreased production of pro-inflammatory cytokines by cultured leukocytes in response to LPS following a first exposure to the same stimulus. Gamma interferon (IFNγ) and granulocyte monocyte-colony stimulating factor (GM-CSF) are immunostimulatory cytokines that prime monocytes and prevent endotoxin tolerance. In monocytes pretreated with IFNγ or GM-CSF, IRAK-1 expression is upregulated. IFNγ and GM-CSF prevent endotoxin tolerance induced by low doses of LPS, by inhibiting IRAK-1 degradation and by promoting its association with MyD88 after a second LPS stimulation, which in turn leads to NF-κB activation and TNF production (Adib-Conquy & Cavaillon, J. Biol. Chem. 2002, 277: 27927).
Endotoxin tolerance has been defined and analyzed either entirely in vivo or entirely in vitro. In contrast, the hyporeactivity of circulating leukocytes reported in patients with sepsis and often referred to an endotoxin tolerance phenomenon, is an ex vivo observation. Therefore, our objective was to set up an ex vivo model of endotoxin tolerance. Mice were injected intravenously with LPS and their leukocytes derived from different compartments were challenged in vitro with LPS or HKSA for production of TNF. A rapid (1-3h) and dramatic decrease in the production of TNF in response to LPS was observed with circulating leukocytes, splenocytes, peritoneal cells and bone-marrow cells after LPS injection. In contrast, LPS-induced TNF production by bronchoalveolar cells was far less reduced and only very briefly. The kinetics of acquisition of tolerance and recovery were different for the various compartments. "Cross-tolerance" with HKSA did not parallel the phenomenon of endotoxin tolerance as observed with LPS. These data show that endotoxin tolerance, as monitored by ex vivo analysis, is compartmentalized and that bronchoalveolar cells are less likely than peritoneal, splenic or marrow cells to develop endotoxin tolerance (Fitting et al. J. Infect. Dis. 2004, 189, 1295).
6. MODULATION OF INTERLEUKIN-10 PROPERTIES BY ADHERENCE (A-F. Petit-Bertron, M. Adib-Conquy)
We previously demonstrated that adherence is a parameter, which markedly affects the properties of IL-10 on TNF production, TNF mRNA expression and NFκB activation by monocytes/macrophages (Adib-Conquy et al. Int. Immunol. 1999, 11, 689). Tyk2 and STAT3 phosphorylation and suppressor of cytokine signaling (SOCS) 3 expression were induced by IL-10 in human monocytes both in the presence and in the absence of adherence, but a longer activation and/or expression was observed in adherent monocytes. Finally, heme oxygenase-1 (HO-1), an anti-inflammatory molecule, was induced by IL-10 in adherent monocytes, whereas its expression remained low in non-adherent cells (Petit-Bertron et al. J. Leuk. Biol. 2003, 73, 145). Recently, we analyzed by macroarray the effect of IL-10 and adherence on the expression of 1050 genes. Interestingly, IL-10 modulated conversely on Teflon® and plastic the expression of 16 genes, of which SOCS2, SOCS3, coproporphyrinogen oxidase (an enzyme involved in heme biosynthesis)), and several matrix metalloproteinases. We confirmed by RT-PCR the modulation of the SOCS molecules and showed that the expression of the other proteins (except for CD35) was modulated similarly to that of their mRNA. The strong inhibition of coproporphyrinogen oxidase in adherent cells, could be one of the mechanisms by which IL-10 contributes to anemia. This study demonstrates that adherence has a profound modulatory effect on the properties and the signaling induced by IL-10. This observation may partially explain the dual role of this cytokine. We consider that our model can mimic some in vivo events when circulating monocytes encounter IL-10 before adhering to endothelium and marginating within tissues towards inflammatory foci. The macroarray study was done in collaboration with T. Pedron and the plate forme 2 " puces à ADN " from Institut Pasteur (Petit-Bertron et al. Cytokine 2005, 29, 1).
7. SEROTONIN AND CYTOKINE PRODUCTION (I. Cloëz-Tayarani)
Serotonin (5-hydroxytryptamine, 5-HT) is released by activated platelets and can be present at micromolar concentration within the inflammatory site. In order to provide additional insight into the in vivo significance of 5-HT in inflammation, we examined its effects on the production of TNF, IL-1α, IL-1β, IL-6, IL-10 and IL-1ra in LPS-stimulated PBMC. 5-HT inhibited TNF production and increased IL-1β production in PBMC. The inhibitory effect of 5-HT on TNF production was antagonized by ketanserin, a selective 5-HT2A antagonist and mimicked by DOI, a selective 5-HT2A/2C agonist. These findings suggest that the inhibition of TNF production by 5-HT involves the participation of the 5HT2A receptor subtypes in PBMC. Accordingly, we detected the presence of 5-HT2A receptors in PBMC by Western-blots. Our data support a role of 5-HT in inflammation through its effect on cytokine production in PBMC (Cloëz-Tayarani et al. Int. Immunol. 2003, 15 233). More recently we showed an activation of extracellular signal-regulated kinase (ERK) within cultured lymphocytes exposed to 5-HT. This phenomenon seems to be mediated via the 5-HT1A receptor since similar results were obtained with R-(+)-8-hydroxy-DPAT, a selective agonist of 5-HT1A receptor, and 5-HT-induced ERK phosphorylation was inhibited by WAY100635, a selective antagonist of 5-HT1A receptor (Cloëz-Tayarani et al. Life Sci. 2004, 76, 429).
Figure 1. From bacteria to disease. During infection, host cells sense bacterial derived products (lipopolysaccharide, LPS; peptidoglycan, PGN; DNA ] via specific receptors (Toll-like receptors, TLR; peptidoglycan recognition proteins, PGRP; nucleotide-binding oligomerization domain, Nod). The activation of leukocyte results in the production of numerous mediators of innate immunity and inflammation (tumor necrosis factor, TNF; interleukin-1, IL-1; nitric oxide, NO ). Other events triggered by bacterial derived products (complement activation, coagulation) amplify the inflammatory process. An exacerbated production of inflammatory mediators contribute to organ dysfunction and eventually death. Concomitantly, the inflammation process is counter-regulated by the induction of anti-inflammatory mediators (interleukin-10, IL-10; IL-1 receptor antagonist, IL-1ra; soluble TNF receptors, sTNFR ), acute phase proteins (including LPS binding protein, LBP; soluble CD14, sCD14 ), neuropeptides and hormones. These mediators together with the occurrence of apoptosis of leukocytes contribute to an altered immune status (Annane D., Bellissant E., Cavaillon J-M. Septic Shock, The Lancet, 2005, 365, 63).
Figure 2. LPS-induced intracellular signaling cascade. Following its binding to CD14, LPS interacts with MD2 / TLR4 signaling complex. Two main pathways are activated either MyD88-dependent leading to NF-κB activation and IL-1 and TNF production, or TRAM/TICAM-2 and TRIF/TICAM-1-dependent leading to the activation IRF3 and interferon-β (IFNβ) production. Other down-stream pathways lead to the activation of Sp1, a prerequisite for IL-10 production, while pI-3 kinase is involved in IL-1ra induction. Numerous intracellular molecules (MyD88s, Tollip, IRAK-M, SOCS-1, pI-3-kinase) and cell surface molecules (ST2, SIGIRR) down-regulate the activating cascade. During sepsis, leukocyte reprogramming is associated with some altered (grey background), or enhanced (yellow background) cellular signaling pathways that lead to the modified production of TNF, IL-1β IL-1ra and IL-10 in response to TLR4 ligands. This cartoon illustrates our published works (Adib-Conquy et al. Am. J. Respir. Crit Care Med. 2000, 162, 1877, Adib-Conquy et al J. Leuk. Biol. 2001; 70:30; Adib-Conquy et al. Am. J. Respir. Crit Care Med. 2003, 168, 158), unpublished work and other reports (Learn et al. J. Biol. Chem. 2001; 276:20234 ; Solomon et al. J. Clin. Invest. 1998; 102:2019). (adapted from Annane D., Bellissant E., Cavaillon J-M. Septic Shock, The Lancet, 2005, 365, 63).
Keywords: Toll-like receptor, sepsis, intracellular signaling, endotoxin, innate immunity