Nod1 and Nod2 induce CCL5/RANTES through the NF-κB pathway

Authors


Abstract

The Nod-like receptor proteins Nod1 and Nod2 participate in innate immune responses against bacteria through intracellular detection of peptidoglycan, a component of bacterial cell wall. Recent evidence has demonstrated that Nod1 stimulates the release of chemokines that attract neutrophils at the site of infection, such as CXCL8/IL-8 in humans, and CXCL1/keratinocyte-derived chemokine and CXCL2/MIP-2 in mice. We aimed to determine whether Nod proteins could trigger the release of CCL5/RANTES, a chemokine known to attract a number of immune cells, but not neutrophils. Our results demonstrate that activation of both Nod1 and Nod2 results in substantial secretion of CCL5 by murine macrophages. Moreover, in vivo, the intraperitoneal injection of murine Nod1 or Nod2 agonists resulted in a rapid secretion of CCL5 into the bloodstream. We also observed that Nod-dependent secretion of CCL5 did not correlate with the induction of the interferon-β pathway, a major signaling cascade for the activation of CCL5 by viruses. In contrast, we identified a key role of the NF-κB pathway in Nod-dependent stimulation of the CCL5 promoter. Together, these results identify a novel target downstream of Nod1 and Nod2, which is likely to play a key role in orchestrating the global Nod-dependent immune defense during bacterial infections.

Abbreviations:
BMDM:

bone marrow-derived macrophages

DAP:

diaminopimelic acid

HEK:

human embryonic kidney

KC:

keratinocyte-derived chemokine

MDP:

muramyl dipeptide

M-TriDAP:

DAP-containing muramyl tripeptide

NLR:

Nod-like receptors

PRM:

pattern recognition molecules

TRIF:

Toll/IL-1 receptor domain-containing adaptor protein inducing IFN-β

Introduction

Nod1 and Nod2 belong to a family of pattern recognition molecules (PRM) known as Nod-like receptors (NLR) or NACHT-LRR proteins. Members of this family also include NALP proteins, IPAF and NAIP, and recent evidence has demonstrated that some NLR serve as intracellular sensors of bacteria and/or danger signals 13. The role of Nod1 and Nod2 as intracellular surveillance molecules implicated in the innate immune responses to bacteria has been demonstrated in vitro and in vivo. In addition, Nod1 and Nod2 are believed to play a central role in the control of immune homeostasis and inflammation at mucosal surfaces. This assumption originates from the discovery that mutations in Nod2 are associated with susceptibility to Crohn's disease in humans 4, 5. Similarly, mutations in Nod1 have been shown to confer susceptibility to asthma and inflammatory bowel disease 6, 7, even though this last observation remains controversial 8.

Nod1 and Nod2 both detect peptidoglycan sub-structures from bacterial peptidoglycan, a cell wall component of virtually all bacteria 914. While Nod2 detects muramyl dipeptide (MurNAc-L-Ala-D-Glu, MDP) 11, 13, the human form of Nod1 senses a diaminopimelic acid (DAP)-containing muramyl tripeptide (muramyl-L-Ala-D-Glu-mesoDAP, M-TriDAP) 9, 10, and murine Nod1 preferentially detects a DAP-containing muramyl tetrapeptide (muramyl-L-Ala-D-Glu-mesoDAP-D-Ala) 14. Following detection of their respective peptidoglycan-derived structures, Nod1 and Nod2 engage signaling pathways that are distinct from those triggered by Toll-like receptors (TLR), a family of membrane-anchored PRM. In particular, both Nod1 and Nod2 trigger the recruitment of receptor-interacting protein 2 (also known as RICK or Cardiak), resulting in the activation of NF-κB and JNK/SAPK pathways 3. For other signaling pathways, such as those implicating p38, ERK or IFN-regulatory factor-3, the ability of Nod1 or Nod2 to induce them has not been studied in detail.

One way by which Nod proteins Nod1 and Nod2 seem to promote host defense to bacterial pathogens is through the activation of anti-microbial peptides. In vitro, both Nod1 and Nod2 have been shown to trigger the secretion of β-defensin-2 1517. In vivo, the protective role of Nod2 in mice orally infected with Listeria seems to be mediated, at least in part, via the Nod2-dependent induction of intestinal cryptdins 18. Another clear role of Nod proteins in innate immune defense against pathogens is the ability of these proteins to trigger cytokine (IL-1β, IL-6 and TNF) and chemokine [CXCL8/IL-8, CXCL-1/keratinocyte-derived chemokine (KC) and CXCL2/MIP-2] secretion in infected cells 1. Interestingly, IL-8, KC and MIP-2 are chemokines that attract neutrophils at the site of inflammation. In agreement with this assumption, Masumoto et al.19 have recently demonstrated in vivo that intraperitoneal injection of a Nod1-specific agonist results in the recruitment of neutrophils at the injection site.

Aside from the recruitment of polymorphonuclear leukocytes to the site of injury or infection, it is critical that a large repertoire of immune cells can be recruited in order to generate a full-blown response to invading pathogens. In this view, CCL5/RANTES represents an important chemokine to analyze in models of microbial infection. Indeed, CCL5 is potently induced following viral and bacterial infections and is detected by several receptors (CCR1, CCR3, CCR4 and CCR5) at the surface of a number of immune cells that are activated and recruited to the site of infection 20. These cells include macrophages, T cells, eosinophils, natural killer cells, platelets, immature dendritic cells and basophils, but not neutrophils. Therefore, it is of particular interest to determine whether Nod proteins, in addition to their ability to trigger chemokines acting as neutrophils chemoattractants, could also induce, through CCL5 stimulation, a broad immune response to bacterial peptidoglycan.

In this study, we provide evidence that stimulation of Nod1 and Nod2 pathways leads to the secretion of CCL5 by murine macrophages, through pathways that are independent from MyD88 and Toll/IL-1 receptor domain-containing adaptor protein inducing IFN-β (TRIF). Importantly, we observed that, in vivo, the intraperitoneal injection of FK156 or MDP, two specific agonists of murine Nod1 and Nod2, respectively, resulted in secretion of CCL5 into the bloodstream. Next, we turned on to analyze the pathways through which Nod proteins stimulate CCL5 secretion. Our results strongly suggest that the induction of the IFN-β pathway is dispensable for Nod-dependent secretion of CCL5, which is at odds with the mechanism by which viruses induce this chemokine in most cases. In contrast, by analyzing the response of the CCL5 promoter to Nod-dependent activation in epithelial cells and in macrophages, we could identify a key role of the NF-κB pathway in this process. Together, these results identify CCL5 as a key new target downstream of Nod1 and Nod2 activation, and these observations provide further insights into the global role of Nod1 and Nod2 in orchestrating the recruitment of immune cells to fight off bacterial infections.

Results

FK156 and MDP trigger CCL5 secretion in murine macrophages

We stimulated bone marrow-derived macrophages (BMDM) from wild-type (WT), Nod1–/– and MyD88–/– mice with TLR2 (Pam3CysSK4), TLR3 (dsRNA), TLR4 (LPS), TLR9 (CpG DNA), Nod1 (FK156) or Nod2 (MDP) agonists, and collected cell culture supernatants 20 h post-stimulation. We then measured by enzyme-linked immunosorbent assay (ELISA) the secretion of CCL5 and, as a comparison, also determined the secretion of KC, a prototypical neutrophil-attracting chemokine. In agreement with our previous study 14, we noted that FK156 triggered the secretion of KC in a Nod1-dependent manner, and we also observed that all TLR agonists tested could trigger KC secretion (Fig. 1A). Interestingly, TLR-dependent, but also Nod-dependent secretion of KC was found to be MyD88-dependent; in the case of FK156, however, the effect was only partial. It must be noted that MyD88 deficiency impairs the signaling pathways downstream of TLR, but also of the IL-1 receptor. Therefore, it is possible that the partial requirement of MyD88 for Nod-dependent secretion of KC might reflect the existence of a paracrine loop involving the secretion of IL-1β in response to Nod stimulation, followed by the IL-1β-dependent stimulation of KC secretion.

Figure 1.

FK156 and MDP trigger CCL5 secretion in murine macrophages. (A, B) BMDM from (A) C57BL/6J (WT), Nod1–/–, MyD88–/– or (B) TRIF–/– mice were stimulated 20 h with Pam3CysSK4 (Pam3, 1.5 ng/mL), dsRNA (1 µg/mL), MDP (10 µM), FK156 (10 µM), LPS (10 ng/mL), CpG DNA (1 µM). KC and CCL5 in the supernatants were measured by ELISA. NS, non-stimulated. Data show the mean ± SEM of duplicates and are representative of three independent experiments. (C) Mouse peritoneal macrophages from WT and Nod2–/– animals were stimulated overnight with MDP (10 µM), LPS (10 ng/mL) or dsRNA (1 µg/mL), and CCL5 in the supernatants was measured by ELISA. NS, non-stimulated. Data show the mean ± SEM of duplicates and are representative of three independent experiments.

Next, we observed that TLR and Nod agonists also potently triggered the secretion of CCL5. As expected, Nod1 deficiency dramatically affected the response to FK156, but not to Nod2 or TLR agonists. However, in sharp contrast with results obtained for KC, MyD88 was dispensable for TLR3- and TLR4-dependent, but also Nod-dependent, secretion of CCL5, but was found indispensable for TLR2- and TLR9-dependent secretion of CCL5.

In order to get more insights into this last observation, we stimulated BMDM from WT and TRIF-deficient mice with the same TLR and Nod agonists (Fig. 1B). We noted that, while KC secretion was comparable, the secretion of CCL5 in response to dsRNA and LPS was altered in TRIF-deficient mice. Importantly, TRIF deficiency was found not to alter the response to the Nod agonists, MDP and FK156. Finally, in order to confirm that MDP-induced secretion of CCL5 was dependent on Nod2, we analyzed the response of WT and Nod2–/– macrophages to MDP, LPS and dsRNA (Fig. 1C). While Nod2 deficiency had no or marginal effect on LPS- and dsRNA-induced CCL5 secretion, the response to MDP was profoundly altered. This result, together with the data presented in Fig. 1A, demonstrates that the secretion of CCL5 in response to FK156 and MDP fully depends on Nod1 and Nod2, respectively.

Nod signaling is independent of IFN-β in murine BMDM

CCL5 expression is strongly induced upon activation of IFN-α/β, and this pathway is critically involved in the CCL5-dependent response to viruses 21. Therefore, we aimed to characterize, in the conditions in which we observed CCL5 secretion in murine macrophages (see Fig. 1), whether the type I IFN pathway was induced. To monitor this pathway, we measured by ELISA the production of IFN-β in the cell culture supernatants of BMDM (WT and MyD88–/–) stimulated 20 h with Pam3CysSK4, dsRNA, LPS, CpG DNA, FK156 or MDP. In the case of TLR agonists, we noted that Pam3CysSK4 and CpG DNA, even though they could trigger substantial secretion of CCL5 (see Fig. 1A), did not trigger IFN-β (Fig. 2A). As expected, dsRNA and LPS stimulation of BMDM resulted in IFN-β secretion that was independent of MyD88 (Fig. 2A). Interestingly, MDP and FK156 failed to stimulate IFN-β secretion, suggesting that the induction of CCL5 secretion by these agonists observed in murine macrophages (see Fig. 1A) was independent of the type I IFN pathway.

Figure 2.

Nod signaling is independent of IFN-β in murine BMDM. (A) Supernatants of stimulated BMDM from C57BL/6J (WT) and MyD88–/– mice were measured by ELISA for IFN-β secretion. Data show the mean ± SEM of duplicates and are representative of three independent experiments. (B) BMDM from WT or Nod1–/– mice were infected with L. monocytogenes for 2, 4, 6 or 8 h at a multiplicity of infection of 10, and analyzed for expression of IFN-β gene by quantitative real-time PCR. Values were normalized to those obtained with the GAPDH housekeeping gene. To calculate inducibility, the values obtained from uninfected macrophages (time point 0 h) were used as a calibrator. Data show the mean ± SEM of duplicates and are representative of three independent experiments.

Nod proteins are intracellular sensors of bacteria, and it is well characterized that infection of macrophages or epithelial cells with Listeria monocytogenes triggers a potent induction of the type I IFN pathway 22. Therefore, we next infected WT and Nod1–/– BMDM with L. monocytogenes for up to 8 h and monitored the resulting induction of IFN-β expression by real-time PCR. Our results clearly establish that Nod1 is dispensable for Listeria-dependent IFN-β induction (Fig. 2B). Since we have already demonstrated that, in similar conditions, Nod2 was also not implicated in this pathway 23, we concluded that Nod1 or Nod2 are unlikely to contribute significantly to bacteria-dependent stimulation of type I IFN responses.

FK156 and MDP trigger CCL5 secretion in vivo

Next, we aimed to investigate whether Nod agonists could trigger CCL5 in vivo. To do so, we directly injected FK156 or MDP into the peritoneal cavity of WT, Nod1–/– and Nod2–/– mice, according to procedures that we have set up previously 14. Two hours post-injection, sera were collected and cytokines measured by ELISA. While KC was considerably induced by injection of FK156 or MDP, as we have described previously 14, we observed that the Nod agonists also induced the release of CCL5 into the bloodstream of the injected animals, albeit more modestly than KC (Fig. 3). As expected, FK156 and MDP failed to trigger CCL5 release from Nod1–/– and Nod2–/– animals, respectively. Interestingly, we could not detect significant accumulation of IFN-β in the sera of the injected animals, again arguing for a minor, if any, contribution of the type I IFN pathways in Nod-dependent induction of CCL5. It must be noted that, in contrast to Nod agonists, intraperitoneal injection of LPS resulted in a significant accumulation of IFN-β in the sera of the injected animals (data not shown).

Figure 3.

FK156 and MDP trigger CCL5 secretion in vivo. WT, Nod1–/– and Nod2–/– mice were injected intraperitoneally with 80 μg FK156 or MDP in 500 μL of endotoxin-free water, and blood was collected 2 h after inoculation. KC, CCL5 and IFN-β were measured in the serum by ELISA. For each condition, three mice were injected and the values represent the mean of individual measurements. The experiment was performed three times with similar results. For the statistical analyses, Nod1–/– and Nod2–/– groups were compared to WT mice for each stimulation; *p<0.05, **p<0.02.

Nod signaling triggers NF-κB but not type I interferon pathway in vitro

In order to get further insights into the signaling pathways through which Nod1 and Nod2 could stimulate CCL5 secretion, we first evaluated the capacity of Nods to turn on type I IFN pathways in vitro, and aimed to compare this to their well-described ability to potentiate the NF-κB pathway. To do so, human embryonic kidney (HEK)293T cells were over-expressed with Nod1 or Nod2 expression vectors, in the presence or absence of their respective peptidoglycan agonist, and the induction of signaling pathways was monitored by measuring the activation of pathway-specific luciferase-reporter constructs. While Nod1 and Nod2 activation resulted in strong activation of NF-κB, they failed to stimulate the type I IFN pathway, as observed using two distinct luciferase-reporter constructs (pISRE-Luc and pIFN-β-Luc) (Fig. 4A). Of note, TBK1 over-expression was used in these experiments as a positive control for both pISRE-Luc and pIFN-β-Luc reporter constructs.

Figure 4.

Nod signaling triggers NF-κB but not type I IFN pathway in vitro. (A) Human HEK293T cells were transiently co-transfected with 1 ng Nod1 or Nod2 expression plasmids together with NF-κB, ISRE or IFN-β luciferase-reporter constructs, and β-gal construct. When indicated, M-TriDAP (250 nM) and MDP (250 nM) were co-transfected along with Nod1 or Nod2, respectively. For each well, raw value was normalized to the β-galactosidase activity. Co-transfection with 50 ng TBK1 expression plasmid was used as a positive control for stimulation of ISRE and IFN-β constructs. Results are presented as relative light units (RLU) and data show the mean ± SEM of duplicates and are representative of three independent experiments. (B) RAW264.7 cells were transiently transfected with 50 ng Nod1 or Nod2 expression plasmids together with NF-κB or ISRE luciferase-reporter constructs, and β-gal construct. For each well, raw value was normalized to the β-galactosidase activity. Stimulation of cells with either LPS (100 ng/mL, 6 h) or dsRNA (1 μg/mL, 6 h) was used as a positive control for stimulation of ISRE and NF-κB constructs. Data represent the mean ± SE from three independent experiments; *p<0.05, **p<0.02.

In order to demonstrate that the defective activation of type I IFN pathways by Nod1 and Nod2 was not specific to the HEK293T epithelial cell line, we also transfected the murine macrophage cell line RAW264.7 with Nod1 or Nod2 expression vectors, and followed the activation of the NF-κB and type I IFN pathways, using pNF-κB-Luc and pISRE-Luc reporter constructs, respectively (Fig. 4B). As for HEK293T cells, over-expression of Nod1 or Nod2 in RAW264.7 cells resulted in the activation of the NF-κB cascade, but not of the type I IFN pathway, while LPS and dsRNA were potent inducers of both signaling cascades. Therefore, from our experiments it seems unlikely that Nod signaling results in a substantial induction of type I IFN in epithelial or macrophage cell lines in vitro. This is in good agreement with the results presented above either in vivo, or following ex vivo stimulation of murine macrophages.

The NF-κB pathway is indispensable for Nod-dependent activation of the CCL5 promoter

Since Nod1 and Nod2 are well-characterized inducers of the NF-κB pathway, we aimed to determine the contribution of this signaling cascade in the specific activation of the CCL5 promoter. First, we transfected the expression vectors of either Nod1 or Nod2 along with a reporter construct consisting of the region –979 to +8 of the murine CCL5 promoter upstream of the luciferase, in RAW264.7 macrophages. The over-expression of Nod2, and to a lesser extent Nod1, resulted in the activation of the CCL5 promoter (Fig. 5A). Next, we performed similar experiments with a CCL5 promoter in which the critical NF-κB binding site had been disrupted by targeted mutagenesis (CCL-MUT). Strikingly, in addition to the expected reduction in the basal level of promoter activation, we noticed that Nod1 and Nod2 over-expression failed to trigger the mutated promoter (Fig. 5A), therefore demonstrating that NF-κB activation is crucial for Nod-dependent activation of the CCL5 promoter in murine macrophages.

Figure 5.

The NF-κB pathway is indispensable for Nod-dependent activation of the CCL5 promoter. (A) RAW264.7 cells were transiently transfected with WT mouse CCL5 promoter-luciferase (CCL5-WT) and NF-κB mutant promoter (CCL5-MUT) constructs without or with Nod1 or Nod2 expression vectors (effector) or their control vector (pCDNA3). Data represent the mean ± SE from five independent experiments; *p<0.05. (B) Nuclear extracts were isolated from RAW264.7 cells, stably transfected with either pcDNA3 or human Nod2, following MDP or LPS stimulation for 5 h. (Left panel) EMSA was performed with 10 μg of nuclear extract for each sample and a ds oligonucleotide probe containing the –99/–67 region of the mouse CCL5 promoter. (Right panel) Nuclear extracts from RAW264.7 cells stably transfected with hNod2 and stimulated with LPS for 5 h were incubated with the CCL5 promoter probe, together with PBS or various probes, as indicated. NF-κB cs, NF-κB consensus sequence; Mut, mutated NF-κB binding site; Med, medium; FP, free probe. (C) Nuclear extracts from RAW264.7 cells stably transfected with hNod2 and stimulated with LPS for 5 h were incubated with the ds oligonucleotide probe containing the –99/–67 region of the mouse CCL5 promoter, together with isotype control immunoglobulins (IgG), or antibodies against different NF-κB subunits, as indicated.

Finally, we purified nuclear protein extracts from RAW264.7 cells stably transfected with either an empty vector (RAW-pcDNA3) or the expression vector encoding for the human form of Nod2 (RAW-hNod2), which were left either untreated or stimulated with MDP or LPS. These nuclear protein extracts were then incubated together with a ds oligonucleotide probe corresponding to the CCL5 promoter region spanning the NF-κB binding site, and the binding of NF-κB to its site was evaluated by electrophoretic mobility shift assay (EMSA). The probe also contains a GGAA motif, which is constitutively occupied by the transcription factor PU.1 24, and therefore the presence of PU.1 on its site represents an internal normalization control for the EMSA.

We noticed that stimulation of RAW-pcDNA3 cells with MDP or LPS alone resulted in a weak binding of NF-κB complexes to the NF-κB site on the CCL5 promoter (Fig. 5B, left panel). However, interestingly, the binding of NF-κB complexes was substantially increased in cells over-expressing Nod2 and stimulated with LPS, suggesting that a synergy between LPS-dependent and Nod2-triggered pathways results in enhanced activation of NF-κB on the CCL5 promoter (Fig. 5B, left panel). Of note, this result is consistent with the observation that, in murine BMDM, LPS and MDP act in synergy to trigger enhanced secretion of CCL5 (data not shown).

The necessary controls were performed to definitively establish that NF-κB complexes were actually responsible for the slower migration of the probe in the case of RAW-hNod2 stimulated with LPS. First, competition assays were performed using unlabeled probes containing either the NF-κB consensus sequence, a mutated NF-κB site or an unrelated sequence (Fig. 5B, right panel). Second, the oligonucleotide probe and the nuclear proteins from RAW-hNod2 stimulated with LPS were incubated with antibodies against different NF-κB subunits (p65, p50 or c-Rel) (Fig. 5C). Interestingly, only antibodies against c-Rel could efficiently neutralize the binding of NF-κB onto the oligonuclotide probe, thus showing that a substantial fraction of NF-κB complexes interacting with the NF-κB binding site of the CCL5 promoter contain the c-Rel subunit (Fig. 5C). Together, our results demonstrate that, while the type I IFN pathway seems to be dispensable for the subsequent induction of CCL5, the activation of NF-κB by Nod1 and Nod2 appears to represent a critical event for the induction of this chemokine.

Discussion

The Nod proteins Nod1 and Nod2 have recently emerged as key PRM involved in immune responses to bacterial products or infections. Most evidence reported so far argues for a critical role of these proteins in the activation of the NF-κB pathway 3. In line with this assumption, we recently demonstrated that intraperitoneal injection of FK156 results in a strong and fast Nod1-dependent release of KC 14, the induction of which is mainly driven by NF-κB. In parallel, through a global microarray analysis of Nod1-induced genes, Masumoto et al.19 have identified that in epithelial cells, the majority of the genes induced by their Nod1 agonist were known to be up-regulated through NF-κB and pro-inflammatory molecules. Interestingly, the authors of this study concluded that the Nod1-dependent induction of KC and MIP-2 participates in the recruitment of neutrophils.

In the present study, we aimed to broaden our understanding of the role of Nod proteins by analyzing whether the activation of these PRM could induce the secretion of CCL5/RANTES, a chemokine known to be chemotactic for several immune cell populations, including macrophages, T cells, eosinophils, natural killer cells, platelets, immature dendritic cells and basophils, but not neutrophils. Our results clearly establish that FK156 and MDP trigger CCL5 secretion through Nod1 and Nod2, respectively.

These results were obtained in primary murine macrophages, and also in vivo following direct injection of Nod agonists in the peritoneal cavity. It must be noted that the levels of CCL5 measured in the blood following injection remained below the levels of KC (∼400 pg/mL versus ∼20 000 pg/mL, respectively; see Fig. 3); however, they were comparable to the amounts of TNF or IL-6 that we previously reported 14. By contrast, in isolated macrophages, CCL5 secretion induced by Nod agonists was comparable or higher than that of KC. This may reflect the fact that, in the peritoneal cavity, other immune cells than macrophages and non-immune cells such as stromal cells could account for most of the rapid release of KC into the bloodstream, in response to Nod activation. This question needs to be further addressed in the future.

Since CCL5 is known to be turned on by both NF-κB- and type I IFN-dependent pathways, we explored which pathways contribute, downstream of Nod1 and Nod2, to the induction of CCL5. While our results strongly argue for a critical role of NF-κB activation in Nod-dependent induction of the CCL5 promoter, we failed to identify a direct link between Nod signaling and the induction of the type I IFN pathway, which could mediate NF-κB-independent stimulation of CCL5 synthesis. On the basis of our results, we cannot rule out the possibility that a minimal activation of the type I IFN pathway by Nod proteins could, in synergy with NF-κB signaling, result in a robust activation of CCL5.

It is interesting to note that, in the case of viral infections, it has been clearly shown that the induction of the type I IFN responses play a key role in the induction of CCL5. Recent evidence has demonstrated that the engagement of TLR3, RIG-I or Mda-5 upon detection of viral RNA represents the upstream events of the activation of the type I IFN cascade, which is critical for host antiviral responses. In the case of intracellular bacteria, this signaling pathway appears to be mediated by an unknown receptor, which detects specifically bacterial dsDNA 25, 26. The results presented here argue that Nod proteins do not significantly influence the activation of the type I IFN pathway mediated by this unidentified bacterial DNA sensor. This is in agreement with a previous study which demonstrated that intracellular detection of invasive bacteria results in the activation of the type I IFN responses, independently of receptor-interacting protein 2, a critical adaptor protein acting downstream of Nod1 and Nod2 27.

The fact that Nod proteins are implicated in the induction of CCL5, a chemokine that attracts a broad repertoire of immune cells, suggests that these PRM could play an important role in vivo, in the course of a bacterial infection. Because CCL5 recruits immature dendritic cells and T lymphocytes at the site of injury, Nod proteins could then participate not only in the innate immune response to bacterial pathogens, but also orchestrate the activation of adaptive immune responses. In agreement with this assumption, we have demonstrated recently that Nod1 plays a key role in the onset of adaptive immune responses following detection of bacterial peptidoglycan 28. More generally, it will be interesting to analyze both CCL5 induction and local recruitment of immune cells following bacterial challenges, using double-deficient MyD88/TRIF- and Nod1/Nod2-knockout mice. Only then we would be able to estimate the relative contribution of TLR versus Nod proteins (and possibly other NLR) in the global establishment of a cytokine/chemokine network that is critical for the clearance of invading bacterial pathogens. Further investigation will also be required to identify, beyond CCL5, the repertoire of chemokines of the C, CC, CXC and CX3C families that are triggered by Nod stimulation, in myeloid and non-myeloid cells.

In the past three decades, a large literature has demonstrated that muramyl peptides display a number of antiviral activities in vitro and in vivo, against microbes such as influenza, herpes simplex or human immunodeficiency viruses 2931. However, the mechanisms by which these muramyl peptides display antiviral properties remain poorly understood. The observations presented in this study identify CCL5 as a new downstream target of muramyl peptides, which mediate their action through the activation of Nod1 and Nod2. Given the central role played by CCL5 in antiviral responses, our results provide new insights into the mechanisms through which muramyl peptides could act as immunomodulators in immunity against virus infections. More generally, the recent interest for Nod proteins in immunity will certainly provide the acceleration force to revisit the role of muramyl peptides in antiviral therapies, through research in their adjuvant or anti-infectious activities.

Materials and methods

Reagents

Endotoxin-free fetal calf serum (FCS) was from Hyclone (Logan, UT), and used after heat inactivation at 56°C for 30 min. Media, glutamine and antibiotics were from Invitrogen Gibco (Cergy, France). TLR2 agonist Pam3CysSK4 was from EMC Microcollections (Tuebingen, Germany). Escherichia coli 0114:B4 LPS was from InvivoGen (San Diego, CA) and repurified according to the Hirschfeld's procedure 32. MDP was from Calbiochem (Leicestershire, UK). M-TriDAP was from Dominique Mengin-Lecreulx (Orsay, France). Poly(I:C) (dsRNA) was from InvivoGen. CpG (1668) was from Eurogentec (Angers, France). All agonists and media were controlled for lack of endotoxin content by the Limulus amoebocyte gelation activity test performed using the QCL-1000 from BioWhittaker (Verviers, Belgium), according to the manufacturer's recommendations.

Plasmids

Murine CCL5 promoter that extended from –979 to +8 was amplified by PCR with genomic DNA extracted from the spleen of WT C57BL/6 mice as we previously described 33. Luciferase reporter plasmid pNF-κB-Luc was purchased from Stratagene (La Jolla, CA), IFN-β-Luc or ISRE-Luc reporter constructs, as well as positive control plasmid expressing TBK1 (pTBK1) were kindly given by Eliane Meurs (Institut Pasteur). Expression vector for Nod1 and Nod2 were a kind gift from Dr. Nunez (University of Michigan, Ann Arbor, MI). Control plasmid pcDNA3.1 was from Invitrogen (Cergy, France) and β-gal (pCMV-β-gal) was from Clontech (Palo Alto, CA). All plasmid DNA for transfection were prepared with Qiagen Endo-free Maxi-Prep kits (Qiagen Inc., Valencia, CA).

Mice

Female mice (6–10 wk old) were used for this study. C57BL6/J mice were purchased from Janvier (Le Genest, France) and were used as control mice. Mice deficient for Nod1 (Nod1–/–) or Nod2 (Nod2–/–), respectively given by John Bertin (Millenium, Cambridge, MA) and Jean-Pierre Hugot (Hôpital Robert Debré, Paris, France), have been further backcrossed eight times into C57BL6/J mice. Lps2 (TRIF–/–) mice were kindly given by Dr. Beutler (The Scripps Research Institute, La Jolla, CA). MyD88-deficient mice were provided by Shizuo Akira (University of Osaka, Osaka, Japan). All protocols were reviewed by the Institut Pasteur competent authority for compliance with the French and European regulations on Animal Welfare and with Public Health Service recommendations, and by the Animal Care Committee of the University of Toronto.

Macrophage stimulation

Macrophages derived from bone marrow cells of femur and tibia of two mice were obtained as followed: Five millions leukocytes from filtered flushed cells were plated in 100-mm culture dishes from TPP (Trasadingen, Switzerland) in 10 mL of complete medium with RPMI/10% FCS, antibiotics and 10% L929 supernatant. After 3 days of culture (5% CO2, 37°C), 10 mL of complete medium was added and at day 7, adherent cells were removed with PBS-EDTA treatment, counted and plated in 96-well plates (300 000 cells in 100 µL per well) and used as BMDM.

Mouse peritoneal macrophages were elicited by injection of 1.5 mL of thioglycolate medium from Bio-Rad (Marnes La Coquette, France) in the peritoneal cavity 4 days before peritoneal lavage with 5 mL of PBS complemented with Heparin Choay (10 U/mL) from Sanofi (Gentilly, France). Cells were pooled from five to six mice, then centrifuged (256 × g, 5 min, room temperature), suspended to 500 000 cells/200 µL in RPMI/3% FCS and seeded in 96-well plates. After 90 min of incubation (37°C, 5% CO2), cells were thoroughly washed with PBS to remove non-adherent cells and 100 µL of RPMI/10% FCS/100 U/mL penicillin/100 µg/mL streptomycin was added.

BMDM and peritoneal macrophages were left at least 2 h before being stimulated in duplicates or triplicates. Cells were pretreated 30 min before stimulation with cytochalasine D from Calbiochem (UK) to a final concentration of 1 µM, then stimulated by addition of 100 µL of complete medium with the agonists. After 20 h of stimulation, supernatants were aliquoted and frozen at –20°C for subsequent cytokine dosage.

In vivo injection of muropeptides

Mice were injected intraperitoneally with 80 µg of muropeptide (MDP or FK156) in 500 µL of endotoxin-free water, as previously described 14. Blood was collected 2 h after inoculation, and sera were frozen for subsequent cytokine dosage.

Cytokine dosage

The concentrations of murine cytokines (KC, IL-6, RANTES) released into the medium were measured using duoset ELISA kits from R&D (Lille, France), and from PBL Laboratories (Piscataway, NJ) for IFN-β secretion.

Transfection and luciferase reporter assays in HEK293T cells

HEK293T cells were from ATCC and grown in DMEM with 10 mM HEPES, supplemented with 2 mM glutamine and 10% FCS. HEK293T cells were seeded in 24-well plates (105 cells/well). The following day, cells (reaching 30–50% of confluence) were transiently transfected with a luciferase-reporter construct along with constructs expressing Nod1 or Nod2 using the Fugene 6 reagent from Roche Diagnostics (Mannheim, Germany), according to the manufacturer's recommendations. Briefly, 75 ng of luciferase-reporter construct was co-transfected with 1 ng of Nod1 or Nod2 constructs and 20 ng of β-galactosidase expression plasmid. For each transfection point, total DNA was adjusted to 300 ng by empty vector pcDNA3.1. For experiments with IFN-β or ISRE promoter constructs, a construct expressing TBK1 was co-transfected as a positive control.

After 20 h of transfection, cells were stimulated for 6 h in triplicate. Then supernatants were discarded and cells were lysed in 100 µL of lysis buffer (25 mM Tris pH 8, 8 mM MgCl2, 1% Triton X-100, 15% glycerol, 1 mM DTT) for 5 min at room temperature. Lysed cells (10 µL) were analyzed for luciferase activity in lysis buffer complemented with 2 mM luciferin and 1 mM ATP, from Sigma Aldrich (St Quentin, France), using a Microlumat plus from Berthold Technologies (Bad Wildbad, Germany). For each well, relative luciferase activity was normalized to β-galactosidase activity, in 10 µL of cell lysates. Results are expressed as the mean ± SD of triplicates.

Transfection and assay of RAW264.7 cells

Transient transfections were performed by electroporation essentially as previously described 34. Briefly, murine macrophages of cell line RAW264.7 were grown to confluence (106/mL) in RPMI 1640 medium supplemented with 10% FCS. Cells were transfected with mCCL5 promoter-, NF-κB- and ISRE-luciferase-reporter constructs. Cells were collected, washed once with RPMI 1640 medium, and resuspended in the same medium at a concentration of 10×106 cells/400 µL. Then 400 µL of cell suspension and 40 μL of DNA (6 μg of luciferase-driven reporter gene, 6 μg of CMV-β-galactosidase, 10 μg of pBluscriptlIKS+ or pcDNA3) were placed in 0.45-cm electroporation cuvettes (Gene Pulser; Bio-Rad Laboratories, Richmond, CA) and electroporation was carried out at 975 μFD and 300 V. Transfected cells were collected and resuspended to 1×106/mL in RPMI 1640 plus 10% FCS, and chloroquine was added to a final concentration of 10 μM. Cells were placed in wells (2 mL/well) of a 24-well plate and incubated for 16 h at 37°C in a 5% CO2 atmosphere. Cells were treated with murine rIFN-γ (10 ng/mL) and LPS (1 μg/mL). After 7 h stimulation cells were harvested and lysed by Triton. Lysates were used for both luciferase and β-galactosidase assays.

Bacterial infections

L. monocytogenes strain EGDe was grown and used to infect BMDM as described 23.

EMSA

EMSA and supershifts were performed as described previously 35, and the oligonucleotide probe spanning the NF-κB and PU.1 sites on the CCL5 promoter have been previously described 24. Competitive EMSA was performed by incubating nuclear extracts isolated from RAW cells stably transfected with hNod2 (RAW-hNod2) and treated by LPS, with the –99/–67 probe and various competitors as indicated. A “supershift” EMSA was performed with the –99/–67 probe and the nuclear extract used in the competition EMSA with various anti-NF-κB antibodies, as indicated, following procedures previously described 24.

Statistical analysis

Unless indicated in the figure legend, the results are given as mean SEM. Statistical analysis was performed with Graphpad Prism software using a Student's t-test. A p value <0.05 was considered significant.

Acknowledgements

L.L.B. acknowledges a fellowship by the Association François Aupetit (AFA) followed by a ‘Bourse fin de thèse’ from the French Fondation Recherche Médicale (FRM). J.H.F. is supported by an Erwin Schrödinger Research Fellowship (J2630-B13) of the Austrian Science Fund (FWF). We thank Millenium Pharmaceuticals for providing us the Nod1–/– mice, Drs. J. P. Hugot and M. Giovannini for the Nod2–/– mice, and Dr. B. Beutler for the TRIF-deficient mice. D.J.P. is supported by grants from the Canadian Institutes for Health Research and Howard Hughes Medical Institutes. Research in the laboratory of SEG is supported by grants from the Canadian Institutes for Health Research and the Crohn's and Colitis Foundation of Canada.

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

Ancillary