Report

Modulation of Gene Expression via Disruption of NF-κB Signaling by a Bacterial Small Molecule

See allHide authors and affiliations

Science  11 Jul 2008:
Vol. 321, Issue 5886, pp. 259-263
DOI: 10.1126/science.1156499

Abstract

The control of innate immune responses through activation of the nuclear transcription factor NF-κB is essential for the elimination of invading microbial pathogens. We showed that the bacterial N-(3-oxo-dodecanoyl) homoserine lactone (C12) selectively impairs the regulation of NF-κB functions in activated mammalian cells. The consequence is specific repression of stimulus-mediated induction of NF-κB–responsive genes encoding inflammatory cytokines and other immune regulators. These findings uncover a strategy by which C12-producing opportunistic pathogens, such as Pseudomonas aeruginosa, attenuate the innate immune system to establish and maintain local persistent infection in humans, for example, in cystic fibrosis patients.

The innate immune system is activated in response to invading microbial pathogens through evolutionary conserved receptor-dependent mechanisms (1). For example, in mammals, the Toll-like receptor 4 (TLR4) recognizes lipopolysaccharide (LPS) as a generic signal for an infection by Gram-negative bacteria (2). This in turn leads to the rapid activation of the nuclear transcription factor NF-κB and the expression of genes encoding proinflammatory cytokines, including tumor necrosis factor–α (TNF-α). Upon cellular stimulation with TNF through its receptors (TNFR), a positive signaling feedback loop in the NF-κB pathway prolongs LPS-induced gene expression (3, 4). NF-κB–dependent processes, in concert with other signaling cascades such as the p38 protein kinase pathway (5, 6), result in coordinated physiological responses that are critical for pathogen elimination (710). Despite the activation of the innate immune system, highly virulent bacteria are able to cause acute severe disease resulting from extensive bacteremia, and some opportunistic pathogens have evolved mechanisms to establish persistent infections.

We hypothesized that the distinct pathogenesis of highly virulent and opportunistic pathogens is the result of differences in their ability to affect signaling in macrophages induced through TLR-responsive pathways. The Gram-negative bacterium Salmonella typhimurium and the Gram-positive bacterium Staphylococcus aureus were chosen as representative pathogens that activate cells mainly through TLR4 and TLR2, respectively. As a classic example of an opportunistic pathogen, the Gram-negative bacterium Pseudomonas aeruginosa, which causes persistent infections in humans, especially in patients with cystic fibrosis (CF), was selected (11, 12). To initiate our investigation, we exposed bone marrow–derived macrophages (BMDMs) to either S. typhimurium, S. aureus, or P. aeruginosa (13), and the macrophage responsiveness was compared by Western blot analysis for the inhibitor of NF-κB (IκB) alpha (IκBα) protein degradation and resynthesis, a distinct feature of NF-κB signaling (14), and the phosphorylation of p38 (p-p38), an additional marker of TLR pathways. The stimulation with these bacteria resulted in similar p-p38 amounts. Also, the macrophage activation in response to S. typhimurium and S. aureus was evident in comparable temporal patterns of IκBα protein degradation and resynthesis; however, we observed a substantial delay in IκBα resynthesis when the cells were treated with P. aeruginosa (Fig. 1A).

Fig. 1.

C12-producing P. aeruginosa alters inducible NF-κB signaling. (A) Comparison of the macrophage responsiveness to P. aeruginosa (P. a.), S. typhimurium (S. t.), and S. aureus (S.a.). Western blot analysis of IκBα, p38, and its phosphorylated form (p-p38) in BMDM extracts after treatment with bacteria are shown. (B) BMDM were stimulated with P. aeruginosa wild type (wt) or lasI mutant (ΔlasI), and total protein extracts were prepared and analyzed as in (A). (C) Chemical structure of C12. (D and E) Western blot analysis of IκBα, p38, and their phosphorylated forms in BMDM extracts (D) or alveolar macrophages (E) after treatment with LPS, C12, or a combination of both.

This difference in NF-κB pathway regulation might be linked to the presence of the bacterial N-(3-oxo-dodecanoyl) homoserine lactone (C12), a product of P. aeruginosa (15), because it was measured in significant concentrations in the samples of P. aeruginosa culture (C12 = 4.7 μM, P < 0.001, n = 5 independent experiments) but not in S. typhimurium or S. aureus samples. We found that the BMDM responses to P. aeruginosa deficient in lasI, the gene responsible for C12 synthesis, showed the normal profile in IκBα degradation and resynthesis, providing evidence that C12, secreted by wild-type P. aeruginosa, caused the observed effect on the NF-κB pathway in macrophages (Fig. 1, A and B, and fig. S1).

C12 is a small bacterial signaling molecule (Fig. 1C) that, in contrast to LPS, activates mammalian cells through TLR4-independent mechanisms (16). In our P. aeruginosa experiments, the macrophages were exposed to a complex mixture of bacterial products containing not only C12 but also LPS. To distinguish whether the observed abnormal NF-κB signaling was mediated by C12 itself or whether it altered LPS response, we investigated the biochemical effects of LPS, C12, and their combination. The temporal profiles of IκBα expression and phosphorylation in response to LPS were substantially impaired in the presence of C12, whereas prolonged p38 phosphorylation was observed (Fig. 1D). The natural stereochemistry of C12 and its structural integrity are required for disruption of LPS-mediated NF-κB signaling (fig. S2). These effects of C12 on LPS-induced signaling were not limited to BMDM, but were also observed in alveolar macrophages (Fig. 1E) and human cells from the THP1 cell line (fig. S3).

The highly coordinated degradation and resynthesis of IκB proteins are critical for stimulus-induced NF-κB activation, whose transcriptional activity is additionally regulated through phosphorylation of its subunit RelA (17, 18). Therefore, we next examined the ability of C12 to modulate LPS-induced phosphorylation of RelA. Because IκBβ plays an important role in LPS-responsive NF-κB regulation (19), especially in the absence of IκBα (20), IκBβ protein levels were also monitored. We found that temporal profiles of RelA phosphorylation, as well as IκBβ degradation in response to LPS, were substantially disrupted in the presence of C12, whereas IκBβ protein levels remained relatively high in cells stimulated with LPS plus C12 (Fig. 2A). Metabolic labeling experiments independently confirmed the effects of C12 on LPS-induced RelA phosphorylation and demonstrated that the phosphorylated form of IκBβ remains bound to RelA in the absence of IκBα (Fig. 2B).

Fig. 2.

C12 disrupts LPS-induced NF-κB signaling. (A) Western blot analysis of IκBβ, RelA, and its phosphorylated form (p-RelA) in BMDM extracts after treatment with LPS, C12, or a combination of both. (B) [32P]Orthophosphate metabolically labeled BMDMs were stimulated with LPS, C12, or a combination of both. The RelA-containing complex was immunoprecipitated with antibodies to RelA and resolved by gel electrophoresis. Phosphoproteins were detected by autoradiography, and their positions on the gel are marked by arrows. (C) The profiles of IKK activity, IκBα, and RelA phosphorylation, as well as IκBα and IκBβ protein concentrations in BMDMs stimulated with LPS alone or in combination with C12. Quantitated and normalized results from three experiments were graphed. (D) BMDMs were pretreated with LPS for 2 hours and then left untreated or stimulated with C12 for 15 or 45 min, as indicated. The obtained extracts were used for immunoblot analysis for IκBα, RelA, CREB, p38, and their phosphorylated forms.

The IκB kinase complex (IKK) phosphorylates IκBα and RelA (21); thus, we examined the IKK activity in macrophages stimulated with LPS, C12, or LPS and C12 together. The results of in vitro kinase assays revealed that although C12 did not induce IKK activity (fig. S4) and modestly reduced the early phase of IKK activation by LPS (fig. S5), the LPS-responsive kinase activity was down-regulated faster in the presence of C12 at late time points, especially after 5 min (Fig. 2C, top). Correlations between the profiles of IKK activity and the corresponding patterns of IκBα and RelA phosphorylation prompted the investigation of whether C12 is able to affect IKK activity in the late stage of LPS stimulation, when newly synthesized IκBα becomes abundant (Fig. 2C and fig. S6). IKK activity was evident by the presence of p-IκBα and p-RelA in cell extracts from macrophages pretreated with LPS for 2 hours, whereas the addition of C12 to these cells resulted in rapid reduction of p-IκBα and p-RelA (Fig. 2D). In contrast, the phosphorylation of p38 and the transcription factor adenosine 3′,5′-monophosphate (cAMP) response element–binding protein (CREB), a target of the p38 pathway (22), was induced in response to C12 (Fig. 2D). These observations support our interpretation that C12 selectively causes abnormal regulation of LPS-induced IKK activity. Although further investigations are required to identify the mechanisms of this regulation, it is unlikely that C12 is a direct inhibitor of IKK activity, because both the purified IKK complex and a constitutive active form of its subunit IKKβ displayed normal kinase activity in the presence of even high doses of C12 (fig. S7).

To define functional consequences of the abnormal NF-κB signaling, we investigated how C12 affects the induction of known NF-κB–responsive genes (table S1) in response to proinflammatory stimuli. To control for specificity, we monitored the mRNA levels of IkBα, IRF-1, and IP-10. The transcription of IkBα is predominantly regulated through the classical NF-κB pathway, whereas the induction of IP-10 and IRF-1 can also be induced independently of NF-κB, such as in response to interferon-γ (IFN-γ) through the Janus kinase/signal transducers and activators of transcription (JAK/STAT) pathway (23). Northern blot analysis revealed that LPS induction of IkBα and other NF-κB–regulated genes in BMDMs was substantially impaired in the presence of C12 (Fig. 3A). Similar results were obtained in lung macrophages and embryonic fibroblasts (fig. S8). This C12 effect was dose-dependent and occurred in cells that remained stimulus-responsive, as evident by the C12-mediated induction of c-jun mRNA (fig. S9). To confirm the specificity of C12-mediated modulation of NF-κB signaling, we measured its effect on IFN-γ–dependent induction of IP-10 and IRF-1 mRNAs. The expression of these mRNAs in response to IFN-γ was unaffected by the presence of C12 (Fig. 3B).

Fig. 3.

C12 specifically impairs inducible NF-κB–regulated gene expression. (A to C) Northern blot analysis of total RNA prepared from BMDMs monitors the effect of C12 on the mRNA expression levels of genes indicated on the right and induced by LPS (A), IFN-γ (B), or HKSA (C). (D) Mouse embryo fibroblasts (MEFs) were stimulated with C12, TNF or a combination of both, and total RNA was prepared. Northern blots showing the mRNA levels of genes are indicated on the right. (E) Western blot analysis of IκBα, IκBβ, RelA, p38, and their phosphorylated forms in extracts prepared from MEFs after treatment with TNF, C12, or a combination of both.

Next, we examined the effect of C12 on the macrophage responsiveness to heat-killed S. aureus (HKSA) containing Gram-positive bacteria-derived TLR ligands. Northern blot analysis demonstrated that C12 substantially impaired HKSA-mediated induction of IkBα and other NF-κB–dependent genes (Fig. 3C). Thus, C12 modulates the ability of distinct TLR ligands to activate NF-κB–dependent gene expression.

To clarify whether C12 targets molecular components of the TLR signaling cascades or the NF-κB pathway itself, we examined the effect of C12 on the induction of NF-κB–regulated genes in response to TNF. Northern blot analysis demonstrated that C12 substantially altered the TNF-mediated induction of IkBα and other NF-κB–dependent genes (Fig. 3D), providing evidence that C12 acts as a general modulator of the NF-κB pathway. Moreover, the patterns of signal processing within the IκB-RelA modules in cells stimulated with TNF or TNF and C12 (Fig. 3E and fig. S10) matched those that were previously observed for C12-mediated effects on the cellular responses to LPS.

To extend our studies, we used transgenic mice harboring a luciferase reporter driven by an NF-κB–responsive promoter of the IkBα or TNF gene. The luminescence of both IκB- and TNF-reporter animals was monitored after injection of vehicle, LPS, C12, or a combination of both. These experiments in mice confirmed our in vitro data that LPS-induced responses were substantially suppressed by C12 (Fig. 4, A and B).

Fig. 4.

C12-mediated inhibition of TNF expression in activated leukocytes could be beneficial for surrounding cells. (A and B) Bioluminescence emission from TNF-reporter (A) or IκB-reporter (B) mice 2 hours after injection with vehicle, LPS, C12, or a combination of both, as indicated. (C) Western blot analysis monitors PARP cleavage as well as the amounts of IκBα and its phosphorylated form in WI38 lung fibroblasts and bronchial epithelial cells stimulated for 2 hours, as indicated. (D and E) Northern blot analysis shows the effect of C12 on rapid LPS-mediated induction of TNF and IκBα mRNAs in primary cells. (F) Inhibitory effect of C12 on LPS-induced TNF production in whole blood.

In TNFR-mediated cellular responses, proper function of the NF-κB pathway is required for protection from TNF cytotoxicity (24); for example, RelA-deficient fibroblasts and macrophages are sensitive to TNF-induced apoptosis (25). The treatment of lung fibroblasts and bronchial epithelial cells with TNF plus C12 resulted in enhanced poly (adenosine 5′-diphosphate—ribose) polymerase (PARP) cleavage (Fig. 4C), which is a biochemical marker indicative of apoptosis (26). As expected, this proapoptotic response was accompanied by the loss of IκBα protein and its phosphorylated form, confirming the abnormal regulation of TNF-mediated NF-κB activation by C12. Macrophages are major producers of TNF that functions in septic shock pathology through proinflammatory autocrine and paracrine mechanisms (27). Therefore, rapid inhibition of TNF expression in LPS-stimulated leukocytes by C12 may indeed be beneficial for the survival of both C12-producing bacterial pathogens and the surrounding cells (Fig. 4, D to F, and fig. S11). This effect might occur in pathological conditions, such as chronic P. aeruginosa colonization of the airways of CF patients or biofilm formation in patients with catheters and on respirators. Further study will determine if these condtions lead to host-cell exposure to local micromolar concentrations of C12 and subsequent effects on host immunity (28).

These findings suggest that, in the case of P. aeruginosa, C12-mediated disruption of NF-κB signaling attenuates TLR4-dependent innate immune responses, thereby potentially promoting persistent infection. The abrogation of LPS-induced NF-κB activity in a variety of mouse strains bearing defects in the TLR4 pathway correlates with the prolongation of some Gram-negative bacterial infections [fig. S12 and (29)]. Therefore, the identification of a mammalian C12 receptor will provide additional insights into how to regulate the interactions between pathogen and host.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1156499/DC1

Materials and Methods

Figs. S1 to S12

Table S1

References

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article