ABSTRACT
Previous studies have indicated that peptidoglycan (PepG) from gram-positive bacteria can exert a priming effect on the innate immune response to lipopolysaccharide (LPS) from gram-negative bacteria. Here, we hypothesized that this priming effect may be preceded by enhanced expression of monocyte CD14, Toll-like receptor 2 (TLR2), and TLR4. In an ex vivo whole human blood model, we observed a substantial synergy between LPS and PepG in the release of tumor necrosis factor alpha and interleukin-1β (IL-1β) over the 24-h experimental period, whereas the effect on IL-8 and IL-10 release was more time dependent. The priming effect of PepG on cytokine release was preceded by a rapid upregulation of CD14, TLR2, and TLR4 expression on monocytes: at 3 hours there was a twofold increase in CD14 expression (P < 0.03), a fivefold increase in TLR2 expression (P < 0.03), and a twofold increase in TLR4 expression (P < 0.03). CD14 and TLR2 remained upregulated throughout the experimental period following exposure to PepG (P < 0.05). Only a transient upregulation of these monocyte receptors was observed following treatment with LPS or LPS plus PepG. In conclusion, the synergistic effect of LPS and PepG on cytokine release is preceded by a reciprocal upregulation of TLR2 and TLR4 by both bacterial cell wall components.
Severe sepsis and septic shock appear to be consequences of a dysregulated innate immune response. It has recently become clear that cytokine production in sepsis is under the control of Toll-like receptor (TLR) signaling. Moreover, synergistic interactions between bacterial components and their activation of TLR signaling have been suggested to contribute to the pathophysiology of sepsis (9, 31, 47). The precise mechanisms underlying the phenomenon of synergy remain unknown.
The innate immune response to bacterial infection is triggered when TLRs located on the cell surface of phagocytes recognize bacterial components, including lipopolysaccharide (LPS) and peptidoglycan (PepG), a major constituent of gram-positive bacterial cell walls. Binding of cell wall components and subsequent activation of intracellular signaling pathways lead to the synthesis and release of proinflammatory cytokines, such as tumor necrosis factor alpha (TNF-α), interleukin-1β (IL-1β), and IL-8, as well as anti-inflammatory cytokines, such as IL-10 (1). Ten mammalian TLRs have been identified, with different ligand specificities (32). Most notably, TLR4 has been identified as the receptor which recognizes bacterial LPS (7, 15, 26, 27) in cooperation with MD-2, a small accessory protein which associates with the extracellular domain of TLR4 (29). In the serum LPS binds to LPS binding protein, which catalyzes its transfer to CD14 on the surface of monocytes (48). CD14 lacks a transmembrane domain, but this binding facilitates the association of a CD14-TLR4-MD2 receptor complex, permitting the induction of intracellular signaling (8, 17). TLR2 recognizes a wide variety of microbial components, including PepG and lipoteichoic acid from gram-positive bacteria (28, 34, 50), and has been shown to be of particular importance in host defense against gram-positive infections (33). PepG has also recently been shown to signal via the intracellular pattern recognition receptor Nod2 (14).
PepG exhibits a number of endotoxic properties (reviewed in reference 40) and was recently demonstrated to cause organ dysfunction when administered to rats (41). An important characteristic of PepG and its structural subunit muramyl dipeptide is their ability to enhance LPS signaling and augment its toxicity (18, 31, 43, 46, 47, 49). This synergistic interaction has been demonstrated in respect to the release of TNF-α and IL-6 in whole human blood (18, 43), the release of proinflammatory cytokines from human monocytic cell lines (46, 49), and the degree of shock and organ injury in a rodent model of sepsis (47). However, the mechanisms underlying this priming effect of PepG on LPS signaling are still not fully understood.
Monocyte expression of CD14 and TLR4 has been shown to be upregulated by exposure to LPS (12, 17, 22), but to date regulation of TLR expression by PepG has not been studied. The aim of this investigation was to determine whether the priming effect of PepG on LPS signaling is associated with upregulation of the expression of the receptors CD14, TLR2, and TLR4 on human monocytes. Additionally, we wished to further characterize the time course of cytokine responses to coadministration of LPS and PepG in comparison to those induced by each bacterial cell wall component alone.
MATERIALS AND METHODS
Bacterial cell wall components.LPS from Escherichia coli 0127:B8 was purchased from Sigma (Poole, Dorset, United Kingdom) and diluted for use in sterile saline (sodium chloride, 0.9% [wt/vol]) (Baxter, Newbury, United Kingdom). PepG was isolated from Staphylococcus aureus as previously described (13). The isolated PepG was enzymatically digested and gave the expected reverse-phase high-pressure liquid chromatography muropeptide profile with no spurious products. Immediately prior to use, PepG was dispersed by sonication (3,000 Hz; three times for 10 s) and diluted in sterile saline. The content of LPS in the PepG preparation was analyzed by the Limulus amoebocyte lysate test (COATEST; Chromogenix, Molndal, Sweden) and was found to be below 2 ng/mg PepG.
Ex vivo whole human blood model.With the approval of the Local Research Ethics Committee, venous blood was collected from healthy volunteers after obtaining written informed consent and was anticoagulated with heparin sodium (CP Pharmaceuticals, Wrexham, United Kingdom) at 30 IU/ml. The blood was incubated in Monovette syringes (Sarstedt, Germany) with gentle agitation at 37°C in the presence of LPS (10 ng/ml), PepG (1 μg/ml), or a combination of the two, as previously described (42). Each experiment was accompanied by a control sample in which only diluent (0.9% saline) was added. At predetermined time intervals (0, 1, 3, 6, 12, and 24 h), blood was removed from the syringes for analysis of cell surface receptor expression by flow cytometry. Plasma was separated and stored at −20°C for subsequent measurement of cytokine levels.
Flow cytometry.One-hundred-microliter aliquots of whole blood were stained with 20 μl of fluorescence-labeled monoclonal antibody or concentration-matched isotype control by incubation on ice for 60 min, protected from light. Red blood cells were then subjected to hypotonic lysis by incubation of samples with FACS lysing solution (Becton Dickinson, Oxford, United Kingdom) for 10 min, and cells were pelleted by centrifugation (1,000 × g for 3 min at 4°C). Cells were washed twice in ice-cold wash buffer (Cell Wash; Becton Dickinson) and finally resuspended in 1% paraformaldehyde (Cell Fix; Becton Dickinson) and then stored protected from light at 4°C until analysis. Antibodies were as follows: fluorescein isothiocyanate (FITC)-conjugated anti-human CD14 (clone 18D11, isotype immunoglobulin G1 [IgG1]) and FITC-conjugated IgG1 isotype control (Diatec, Oslo, Norway), phycoerythrin (PE)-conjugated anti-human TLR2 (clone TL2.1, isotype IgG2a) and PE-conjugated anti-human TLR4 (clone HTA125, isotype IgG2a) (eBioscience), and PE-conjugated IgG2a isotype control (Cymbus Biotechnology, East Leigh, Hampshire, United Kingdom). To enhance identification of the monocyte population for TLR analysis, multicolor flow cytometric analysis was used, with double staining for CD14 and TLR2, CD14 and TLR4, or CD14 and IgG2a isotype control.
Flow cytometry was performed using a single-laser FACScan and CellQuest software (Becton Dickinson). For the analysis of double-stained samples, electronic compensation for fluorochrome spectral overlap was set using appropriate single-stained and unstained samples. Electronic gates were created to define the monocyte population according to their CD14-staining and light scatter characteristics (Fig. 1), and 10,000 monocytes were acquired for each sample. The median fluorescence intensity (MFI) in FL-1 and FL-2 was recorded for each sample, and all samples were analyzed in duplicate. MFI values were corrected for nonspecific binding by subtracting the MFI measured for the matched isotype control sample.
Identification of leukocyte populations during flow cytometry using FL1 versus side scatter dot plot and CD14-FITC-stained whole blood (see Materials and Methods).
Cytokine measurement.Plasma TNF-α, IL-1β, IL-8, and IL-10 were measured by enzyme immunoassay according to the manufacturer's protocol (CLB, Amsterdam, The Netherlands), with all standards and samples run in duplicate.
Statistical evaluation.Data are presented as medians and interquartile ranges. Area-under-the-curve calculations were utilized for quantitative comparison of cytokine production over time. Statistical comparison of two groups was performed using the Wilcoxon signed rank test, while comparison of more than two groups was performed using the Friedman test and the Wilcoxon signed rank test with the SPSS 11.5 software package (Chicago, Ill.). Differences with P values of <0.05 were considered to be statistically significant (n = 6 for all experiments).
RESULTS
Effect of LPS and PepG on release of cytokines.TNF-α, IL-1β, and IL-10 were not detected in plasma from unstimulated control blood at any time. IL-8 could be detected in unstimulated blood from 6 h onwards, but at levels that were significantly lower than those produced by stimulation with any combination of bacterial cell wall components (P < 0.03).
Addition of LPS to whole blood was associated with a rapid, but transient, increase in the plasma TNF-α levels, whereas PepG induced significantly lower TNF-α levels (P < 0.03 for area-under-the-curve comparison) (Fig. 2a). From 3 h onwards, the TNF-α response to combined administration of LPS and PepG was significantly greater than the response to either component alone (P < 0.05) and more than twice the sum of the values obtained with LPS or PepG alone.
Cytokine release in whole blood in response to combined stimulation with LPS and PepG compared to that with each cell wall component alone. Whole blood was incubated at 37°C in the presence of LPS (1 ng/ml) (□), PepG (10 μg/ml) (▵), a combination of LPS and PepG (⧫), or diluent only (0.9% saline) (•). Blood was removed at 0, 1, 3, 6, 12, and 24 h and plasma separated by centrifugation. Plasma cytokine levels were measured by enzyme immunoassay: (a) TNF-α, (b) IL-1β, (c) IL-8, and (d) IL-10. Data are shown as medians and interquartile ranges of values from six donors. *, P < 0.05 for LPS plus PepG compared to LPS alone. #, P < 0.05 for LPS plus PepG compared to PepG alone. □, P < 0.05 for PepG compared to LPS alone.
Compared to TNF-α release, the rise in IL-1β levels induced by LPS was slightly delayed but more sustained (Fig. 2b). As was seen with TNF-α, the IL-1β response to PepG was significantly less than that to LPS (P < 0.03 for area-under-the-curve comparison). From 3 h onwards, coadministration of LPS and PepG was associated with significantly greater (P < 0.05), more-than-additive increases in plasma IL-1β levels than followed stimulation with either cell wall component alone.
IL-8 levels increased progressively in response to LPS throughout the experimental period (Fig. 2c). Although at 1 h the IL-8 response to LPS was greater than that to PepG (P < 0.05), from 6 h onwards exposure to PepG was associated with significantly higher IL-8 levels than was exposure to LPS (P < 0.05). Coadministration of LPS and PepG initially (from 3 to 6 h) resulted in increases in IL-8 levels that were greater than the sum of those seen with LPS or PepG alone (P < 0.05). At 24 h, however, the IL-8 response to PepG was significantly greater than the response to LPS (P < 0.05) and appeared to be greater than the response to PepG combined with LPS (not significant).
The IL-10 response to LPS was delayed, with levels beginning to rise at 6 h and peaking at 12 h (Fig. 2d). PepG administration was associated with much lower IL-10 levels (P < 0.03 for area-under-the-curve comparison), increasing gradually and reaching a maximum at 24 h. There was no evidence of synergy in the IL-10 response to coadministration of LPS and PepG during the first 6 h of stimulation, but beyond this time combined stimulation was associated with significantly higher levels of IL-10 than were seen with either component alone (P < 0.05).
Effect of LPS and PepG on monocyte expression of CD14, TLR2, and TLR4.Exposure to LPS alone resulted in a rapid, transient upregulation in monocyte CD14 expression (Fig. 3a). Exposure to PepG also upregulated CD14, but this effect was more sustained. Combined treatment with LPS and PepG upregulated CD14 expression in a similar manner, but the sustained response seen with exposure to PepG appeared to be abrogated by the coadministration of LPS. At 1 and 3 h, CD14 expression was significantly (P < 0.03) upregulated in response to exposure to LPS, PepG, and LPS plus PepG, but there were no significant differences between the responses to each of the three treatments.
Regulation of monocyte surface expression of (a) CD14, (b) TLR2, and (c) TLR4 in whole blood in response to stimulation with LPS alone (□), PepG alone (▵), and LPS and PepG combined (⧫), compared with unstimulated control blood (•). Whole blood was incubated at 37°C in the presence of LPS (1 ng/ml) and/or PepG (10 μg/ml). Blood was removed at 0, 1, 3, 6, 12, and 24 h, and 100-μl aliquots of whole blood were stained with 20 μl FITC-labeled monoclonal antibody to CD14 and 20 μl PE-labeled monoclonal antibody to TLR2 or TLR4 (see Materials and Methods). Samples were processed by flow cytometry, acquiring 10,000 monocytes per sample, and analyzed using CellQuest software. The MFI for each sample was corrected for nonspecific binding by subtracting the MFI for the concentration-matched isotype control. Data are shown as medians and interquartile ranges of values from six donors. *, P < 0.05 for LPS, PepG, and LPS plus PepG compared to the unstimulated control. #, P < 0.05 for PepG compared to LPS and LPS plus PepG.
TLR2 expression was upregulated by exposure to either LPS or PepG (P < 0.03 at 1 to 12 h), although the response was more sustained following exposure to PepG (Fig. 3b). The response to LPS combined with PepG was similar to that seen with LPS alone. Up to 3 h following exposure to the bacterial components, there was no significant difference between the responses to each of the three treatment conditions, but at 6 and 12 h, TLR2 expression was significantly greater following exposure to PepG than following exposure to LPS or LPS plus PepG (P < 0.05), and TLR2 remained upregulated at 24 h following exposure to PepG.
There was a transient upregulation (1 to 3 h) in TLR4 expression in response to all three treatment conditions (P < 0.03), with no significant difference between the responses and no sustained response to PepG (Fig. 3c).
DISCUSSION
Using a human whole blood model, we have investigated the potential role of differential regulation of the monocyte pattern recognition receptors CD14, TLR2, and TLR4 in the phenomenon of priming by PepG for an enhanced cytokine response to LPS. PepG stimulation was associated with rapid upregulation of CD14, TLR2, and TLR4 on monocytes. Regulation of TLR2 and TLR4 receptor expression by PepG (or its combination with LPS) has not previously been reported for either whole human blood or cultured human monocytes, although TLR2 mRNA has been shown to increase following exposure of cultured human monocytes to LPS or PepG (19). Our observation that exposure to LPS resulted in upregulation in the expression of CD14, TLR2, and TLR4 concurs with previous reports that LPS stimulation of human whole blood leads to upregulation of both TLR2 and TLR4 on monocytes (22). For cultured human monocytes isolated from whole blood, it has been reported that LPS stimulation upregulates surface expression of CD14 (12, 17), TLR2 (12), and TLR4 (17). In contrast, it has also been reported that LPS stimulation of cultured human monocytes results in an increase in TLR4 mRNA expression but a corresponding decrease in its cell surface expression (2). Interestingly we also observed that the expression of each TLR is upregulated not only by its own ligand but also by another bacterial cell wall component which is not thought to associate with the receptor in question. The timing of this receptor upregulation precedes the peak cytokine response by several hours. This reciprocal upregulation of monocyte cell surface expression of TLR2 and TLR4 could increase sensitivity to other ligands and lead to enhanced intracellular signaling and cytokine release.
Other mechanisms related to intracellular signaling pathways may also be involved in the phenomena of priming and synergy. While early investigations suggested that the various members of the TLR family signaled through identical intracellular pathways, recent evidence indicates that, in addition to a common pathway, each TLR has its own independent pathway (reviewed in reference 32) and that the signalings via TLR2 and TLR4 are not equivalent. Several key differences in the intracellular signaling pathways activated by LPS and PepG have been identified (6, 11, 20). A comparison of the signaling molecules activated by TLR2 or TLR4 with those activated in response to concomitant triggering of TLR2 and TLR4 may identify mechanisms downstream of receptor ligation which may be involved in synergy. Other investigators have suggested that upregulation of the signaling molecule MyD88 (49) or the facilitation of translation of TNF-α mRNA (46) may be involved in the synergy between muramyl dipeptide and LPS. Recent research has indicated that CD14 functions as part of a multiligand pattern recognition receptor complex within cholesterol/sphingolipid-rich microdomains (“lipid rafts”) on the monocyte cell membrane (25, 35). TLR4 is recruited to the cluster following binding of LPS (25, 35), which permits the triggering of intracellular signaling cascades. It seems likely that TLR2 may be similarly recruited to the lipid raft cluster. This could provide another mechanism for enhanced receptor interaction and intracellular signaling in response to combined administration of LPS and PepG.
In this study we also describe the IL-1β, IL-8, and IL-10 responses to LPS and PepG over time in a human whole blood model and confirm the effect on TNF-α release (43). In response to costimulation of whole blood with LPS and PepG, we observed an apparently synergistic effect on levels of the proinflammatory cytokines TNF-α and IL-1β in plasma. This may be of relevance to the excess mortality observed in patients with mixed bacterial infections (45). TNF-α is one of the earliest proinflammatory cytokines to be released, appearing in the circulation within an hour of injection of LPS (21). TNF-α recruits and activates neutrophils, macrophages, and lymphocytes and stimulates the release of other cytokines and inflammatory mediators. IL-1β has a spectrum of activity similar to that of TNF-α but tends to be less potent. We also observed a more-than-additive increase in levels of the leukocyte chemotactic factor IL-8 in plasma in response to combined exposure to LPS and PepG, but at later time points PepG stimulation was associated with higher levels of IL-8 than LPS alone or LPS and PepG combined. The extremely high plasma IL-8 levels observed in this whole blood model were consistent with the findings of Wang et al. (44) for isolated monocyte cultures, where they observed significantly higher mRNA and protein expression for chemokines than for cytokines in response to bacterial components, emphasizing the important role of leukocyte recruitment in bacterial infections. The implications of the higher production of IL-8 in response to PepG beyond 6 h are unclear, but it may contribute to the propensity of gram-positive infections to cause abscess (Staphylococcus aureus) and empyema (Streptococcus pneumoniae) formation. The observed lack of synergy in the release of the anti-inflammatory cytokine IL-10 in response to coadministration of LPS and PepG is in keeping with a previous report (43). The enhanced IL-10 response at later time points may be explained by the induction of IL-10 release by paracrine mediators, including TNF-α (37). IL-10 inhibits the release of TNF-α, IL-1, and other proinflammatory cytokines (10) and has been shown to be the functional repressor of monocyte activation in blood from meningococcus patients (3). The clinical implications of these cytokine responses in isolated whole human blood are speculative, but the delayed increase in IL-10 levels following combined administration of LPS and PepG may be permissive for an exaggerated early proinflammatory response and might indicate an early imbalance between pro- and anti-inflammatory cytokines. Excessive production of proinflammatory cytokines, such as TNF-α and IL-1, is associated with poor outcome in sepsis (5, 39). On the other hand, excessive production of anti-inflammatory cytokines, such as IL-10, may also be detrimental due to the development of immunodeficiency (38), and it is likely to be the balance between pro- and anti-inflammatory responses that is important.
The responses we describe may also be relevant to isolated gram-negative infection, since both LPS and PepG are present in gram-negative bacteria. Whole microorganisms comprise multiple components which are capable of activating the innate immune system, and the response to whole bacteria is thought to be a consequence of the cumulative activation of multiple TLRs (23). Whole organisms have been shown to translocate from the gut to the circulation in surgical patients, and this phenomenon is associated with the development of postoperative sepsis (24). Bacterial components, including LPS and PepG, have also been shown to translocate across the gut wall in patients with various pathologies resulting in impaired splanchnic perfusion, including aortic surgery (4), cardiopulmonary bypass (16, 36), and hemorrhagic shock (30). Translocation of bacteria or their components from the gut to the systemic circulation has been postulated as a contributory mechanism in the development of the systemic inflammatory response and multiple organ failure. The enhanced release of proinflammatory mediators in response to coadministration of LPS and PepG may therefore also be relevant to the systemic inflammation associated with bacterial translocation.
Most previous studies in this field have utilized either animal models, which may not be representative of human responses, or cell lines or isolated cell cultures, which incompletely represent events occurring in vivo. The use of an ex vivo whole human blood model to characterize the cellular responses to stimulation with bacterial cell wall components permits ongoing interaction between cells and mediators involved in the innate immune response. It is still unclear how far these in vitro findings can be extrapolated to our understanding of events occurring in vivo in human sepsis. This applies particularly to data from the later time points.
In conclusion, the priming effect of PepG on LPS-induced cytokine release in human blood is preceded by a reciprocal upregulation of TLR2 and TLR4 on the surface of monocytes.
ACKNOWLEDGMENTS
This work was supported by the Joint Research Board of St. Bartholomew's Hospital, HCA International, and the Royal College of Anesthetists.
FOOTNOTES
- Received 15 March 2005.
- Returned for modification 30 May 2005.
- Accepted 28 July 2005.
- Copyright © 2005 American Society for Microbiology
