ABSTRACT
Monocytes are important effector cells in the pathogenesis of bacterial endocarditis since they provide the tissue factor that activates the coagulation system and maintains established vegetations. Monocytes secrete cytokines that can modulate monocyte tissue factor activity (TFA), thereby affecting the formation and maintenance of vegetations. In this study, we show that monocytes cultured for 4 h on a Streptococcus sanguis-infected fibrin matrix mimicking the in vivo vegetational surface express high levels of TFA. This was accompanied by secretion of the proinflammatory cytokines tumor necrosis factor alpha (TNF-α), interleukin-1α (IL-1α), and IL-1β. After a 24-h incubation period the anti-inflammatory cytokine IL-10 could also be detected. Our data show that, whereas TNF-α and IL-1 have a minor role in the induction of TFA by monocytes cultured on a fibrin matrix, TNF-α but not IL-1 plays an important role in the induction of IL-10 by these cells. In turn, our data show that IL-10 is an important factor in the downregulation of monocyte TFA. In summary, we conclude that IL-10 is an important factor in the control of monocyte TFA in endocardial vegetations.
Tissue factor (TF), a transmembrane glycoprotein, is the most important factor in the initiation of the extrinsic coagulation pathway (25). Monocytes can be stimulated by various agents, such as endotoxin, phorbol ester, lipoteichoic acid, and C-reactive protein, to express TF molecules on their membranes, thereby generating TF activity (TFA) (9, 25, 26, 30, 31). We have shown that monocytes also express TFA when they adhere in vitro to fibrin matrixes containing bacteria, in particular bacteria that are known to cause the endovascular disease bacterial endocarditis (BE) (1-4, 38). With respect to the pathogenesis of BE, Drake and coworkers (13) have shown that in vivo TF is a crucial factor in the formation of an endocardial fibrin clot, called endocardial vegetation, in which the infecting microorganisms and blood cells are embedded (33). In a rabbit model of BE it was found that the TF needed to maintain the endocardial vegetation is generated by monocytes that settle from the circulation onto the infected endocardial lesion (2, 3, 6, 38).
Monocyte TFA can be regulated by cytokines. Proinflammatory cytokines, such as tumor necrosis factor alpha (TNF-α) and interleukin-1 (IL-1), induce TFA on monocytes (15, 17, 18, 22, 24, 25). Monocytes isolated from patients with atherosclerosis, a disease in which TNF-α and IL-1 play a significant role, express TF on their membranes (24). Furthermore, macrophages in atherosclerotic plaques express TF (32). Anti-inflammatory cytokines, such as IL-10 and transforming growth factor beta, inhibit or suppress monocyte TFA (14, 18, 20, 24, 29, 30).
It is well known that pro- and anti-inflammatory cytokines each control the other's production (5, 33). TNF-α and IL-1, which are examples of the earliest cytokines produced by monocytes upon appropriate stimulation (15, 28), are able to stimulate monocyte IL-10 production (15, 27, 28, 40). In turn, IL-10 decreases the production of proinflammatory cytokines (11, 12, 15, 39) and ultimately also its own (28). Moreover, IL-10 increases the production of IL-1 receptor antagonist (IL-1ra) and soluble TNF receptors by monocytes, thus hampering the response to IL-1 and TNF-α (15, 16). In vivo these effects might be reflected in patients with BE caused by gram-positive bacteria; some of these patients were found to have circulating TNF-α and, almost invariably, elevated levels of soluble TNF receptor (19). Furthermore, monocytes from patients with Q-fever endocarditis exhibit a marked release of IL-10 and circulating TNF receptor type II and IL-1ra (7, 8).
The present study was undertaken to explore further the role of IL-1, TNF-α, and IL-10 in the induction or amplification of coagulation by fibrin-adherent monocytes. Using an in vitro model of an infected endocardial vegetation, we investigated whether these cytokines are released by surface-adherent monocytes and whether these cytokines modulate the TFA of endocardial vegetations.
MATERIALS AND METHODS
Microorganisms.The dextran-producing Streptococcus sanguis strain (NCTC 7864) was the same as that used in previous studies (2). Streptococci from an overnight culture in Todd-Hewitt broth (Oxoid, London, England) were washed three times in pyrogen-free saline and diluted to a concentration of approximately 108 CFU per ml in RPMI 1640 (ICN Flow Laboratories, Irvine, Scotland) before use in the experiments.
Monocytes.Heparinized buffy coats from 500 ml of peripheral venous blood from different healthy donors were used to isolate monocytes. In short, 500 ml of blood was diluted five times in phosphate-buffered saline (PBS) containing 0.5 U of heparin per ml and layered onto Ficoll-amidotrizoate (P = 1,077 g/ml) (1). After differential centrifugation, the mononuclear cell-rich interface was carefully removed, washed three times in PBS supplemented with heparin, and resuspended in RPMI 1640 to a concentration of 3 × 106 mononuclear cells per ml, one-third of which were monocytes. Cells were incubated overnight in Teflon culture bags (1) at 37°C in a 5% CO2incubator. The next day, the cells were recovered by centrifugation and resuspended in RPMI 1640. For each experiment monocytes from a different donor were used.
Fibrin plates.Fibrinogen (Sigma, St. Louis, Mo.) was dissolved at a concentration of 5 mg/ml in a buffer containing 50 mM triethanolamine (Fluka BioChemika, Buchs, Switzerland) and 100 mM NaCl (Merck, Darmstadt, Germany) at pH 7.45. In a 24-well tissue culture plate (Costar, Cambridge, England), 200 μl of fibrinogen solution and 50 μl of 100 mM CaCl2 (Merck) were mixed in a well. Then 10 μl of thrombin (0.5 U per ml) (Central Laboratory of the Blood Transfusion Service, Amsterdam, The Netherlands) was added to induce fibrin formation. The mixture was allowed to polymerize overnight at 4°C and was used the next day.
Adherence of monocytes to infected or noninfected fibrin matrixes.Approximately 1.5 × 107 CFU ofS. sanguis were layered onto the fibrin matrix and allowed to adhere for 1 h at 37°C and 5% CO2. Then nonadherent bacteria were removed by two consecutive washes with PBS. The mononuclear cell suspension was diluted to approximately 1.5 × 106 monocytes per ml, and 1 ml was layered onto either an infected or a noninfected fibrin matrix. After incubation for 4 h at 37°C and 5% CO2, the TFA of adherent monocytes was determined. In some experiments supernatant was collected for the measurement of cytokine production. An earlier study (1) revealed that equal numbers of monocytes bound to a noninfected orS. sanguis-infected fibrin matrix.
TFA assay.The procedure for determining the TFA of fibrin-adherent monocytes was the same as that used before (1). In short, after 4 h of incubation of the monocytes on fibrin matrixes nonadherent cells were removed by washing with PBS. Adherent cells were then incubated with 125 μl of a buffer containing 0.125 pmol of FVII and 0.125 nmol of CaCl2 for 15 min at 37°C with rotation (40 rpm). Next, 25 μl of FX (10 U per ml) (Kordia, Leiden, The Netherlands) was added. After 5 min a sample of 30 μl was taken from these wells and added to 200 μl of buffer containing EDTA (Boehringer, Mannheim, Germany) to stop FXa formation. Then the mixture was warmed to 37°C. Subsequently, 25 μl of 1-mg-per-ml PefachromeFXa (Kordia), a chromogenic substrate for FXa, was added. After 20 min at 37°C, the conversion of the chromogenic substrate was stopped by the addition of 200 μl of 50% (vol/vol) acetic acid (Merck). The absorption at 405 nm was measured and converted to FXa concentrations as described previously (3). Results are expressed as pmol of FXa/1.5 × 106 monocytes.
Measurement of cytokine concentrations in the supernatants of fibrin-adherent monocytes.The concentration of TNF-α, IL-1α, IL-1β, or IL-10 in the supernatants of monocytes after incubation on noninfected or S. sanguis-infected fibrin matrixes was measured by using a cytokine-specific enzyme-linked immunosorbent assay (ELISA) kit according to the supplier's protocol. ELISA kits for human IL-10 (hIL-10), hIL-1α, and hIL-1β came from the Central Laboratory of the Blood Transfusion Service (35), and the ELISA kit for hTNF-α came from the Biomedical Primate Research Center (Rijswijk, The Netherlands) (35).
Recombinant TNF-α (rTNF-α), rIL-1α, and rIL-10 and specific antibodies against these cytokines were obtained from ITK Diagnostics, Uithoorn, The Netherlands.
Effect of pro- and anti-inflammatory cytokines on monocyte IL-10 release and TFA.Cytokine levels in the supernatants of monocytes were determined after 4 h of culturing on an S. sanguis-infected fibrin matrix. These supernatants were diluted with RPMI 1640 to obtain final concentrations of 0.5 ng of TNF-α per ml, 12 pg of IL-1α per ml, and 93 pg of IL-1β per ml. These cytokine-containing diluted supernatants are referred to as CCS. The effect of TNF-α on the release of IL-10 and TFA was determined by incubation of monocytes with CCS that were preincubated for 30 min with 1 μg of neutralizing polyclonal goat anti-human TNF-α antibody per ml. The effect of IL-1 on IL-10 release and TFA was measured by the incubation of monocytes that had been preincubated for 30 min with 30 ng of IL-1ra per ml with the CCS. After 24 h the supernatants were collected and the IL-10 concentration was determined; TFA was assessed after 4 h of incubation.
Monocytes were incubated on infected as well as noninfected fibrin matrixes in the presence or absence of recombinant cytokines. Cytokines were added at the start of the incubation at concentrations of 0.5 ng of rTNF-α per ml, 5 U of rIL-1α per ml, and 100 U of rIL-10 per ml.
In some experiments rIL-10 was preincubated for 1 h with 50 μg of neutralizing polyclonal goat anti-human IL-10 antibody per ml to check for rIL-10-specific responses.
Statistical analysis.The significance of differences in TFA and cytokine production by monocytes was assessed by means of Student's t test. The significance level was 5%.
RESULTS
TFA and cytokine production of monocytes adherent to noninfected or infected fibrin matrixes.The levels of TFA were low for monocytes incubated for 4 h on noninfected fibrin matrixes (Table1) and were about threefold higher for monocytes incubated on S. sanguis-infected fibrin matrixes. The TFA of monocytes adhering to fibrin was accompanied by TNF-α, IL-1α, and IL-1β production (Table 1) and also, after an incubation period of 24 h, by IL-10 production (Table 1). Similarly, the levels of cytokines produced by incubation of monocytes on infected fibrin were higher than those found for monocytes cultured on noninfected fibrin. This increase in cytokine level was 3.6-fold for TNF-α, 302-fold for IL-1α, 6.3-fold for IL-1β, and 2.3-fold for IL-10 (Table 1). Time course experiments revealed that the levels of TFA and TNF-α were the highest after 4 h of incubation and then declined. The highest level of IL-10 was found after 24 h of incubation of S. sanguis-infected fibrin (Fig.1). The IL-10 produced by the monocytes after adhesion to fibrin matrixes could be the result of the production of TNF-α and IL-1, which are known to induce IL-10 production in monocytes (15, 28, 34). However, the release of IL-10 by monocytes incubated for 24 h on noninfected fibrin in the presence of 0.5 ng of rTNF-α per ml (fold increase of 1.06 ± 0.15; n = 4 ) or 5 U of rIL-1 per ml (fold increase of 1.06 ± 0.56; n = 4 ) was similar to that found for monocytes incubated in the absence of exogenous cytokines. A range of different concentrations of these recombinant cytokines was tested, all of which were without effect (data not shown).
Effect of monocyte adhesion to noninfected or infected fibrin matrixes on TFA and cytokine production
Expression of TFA and production of TNF-α and IL-10 by monocytes bound to S. sanguis-infected fibrin matrixes. Approximately 1.5 × 106 monocytes in 1 ml of medium were incubated on S. sanguis-infected fibrin matrixes for 4 h (white bars) or 24 h (black bars). TFA and cytokine levels were determined as described in Materials and Methods. Results are means ± standard deviations of three (cytokines) or four (TFA) experiments. ∗, P < 0.05 , relative to 4 h of incubation.
Effect of inflammatory mediators on monocyte TFA.The increased levels of TFA that were found for monocytes cultured on anS. sanguis-infected matrix (Table 1) could be the result of the increased endogenous production of TNF-α or IL-1 during the incubation period, since stimulation of monocytes with rTNF-α or rIL-1 can lead to the expression of TFA (17). We therefore assessed the modulatory effect of TNF-α or IL-1α on the level of TFA of monocytes adherent to a noninfected or S. sanguis-infected fibrin matrix. Addition of rTNF-α or rIL-1α to these monocyte cultures, however, did not affect monocyte TFA (Fig.2A). Moreover, neutralization of endogenous TNF-α with a polyclonal goat anti-human TNF-α antibody during monocyte incubation on noninfected fibrin matrixes did not reduce monocyte TFA (data not shown). Furthermore, both cytokines failed to induce TFA on monocytes that were in suspension (data not shown).
Effect of rTNF-α and rIL-1α (A) and rIL-10 (B) on monocyte TFA. Approximately 1.5 × 106 monocytes in 1 ml of medium were incubated for 4 h on either noninfected orS. sanguis-infected fibrin matrixes in the presence or absence of 2 μg of rTNF-α per ml, 80 ng of rIL-1α per ml, 100 U of rIL-10 per ml, or 100 U of IL-10 per ml mixed with 50 μg of neutralizing anti-human IL-10 antibody per ml. Without cytokine addition, monocytes cultured on noninfected fibrin matrixes had a TFA of 32.56 ± 11.26 pmol of FXa/1.5 × 106monocytes (white bar). Results are means ± standard deviations of three experiments and are expressed as the fold increase in FXa formation compared to that of monocytes adherent to noninfected fibrin matrixes in the absence of cytokines (white bar). ∗, P < 0.05 , relative to noninfected fibrin matrix in the absence of cytokines; ∗∗, P < 0.03 , relative to incubation in the absence of IL-10; ∗∗∗, P < 0.05 , relative to incubation in the presence of IL-10.
IL-10 has been shown to downregulate lipopolysaccharide (LPS)-induced monocyte TFA (18). Thus, IL-10 produced by monocytes during adherence to fibrin (Table 1) could be responsible for the decrease in the level of monocyte TFA (Fig. 1). To further explore this, we incubated fibrin-adherent monocytes for 4 h in the presence of rIL-10 or for 24 h in the presence of a neutralizing concentration of a polyclonal goat anti-human IL-10 antibody. Incubation in the presence of rIL-10 resulted in a reduction by approximately 40% of the TFA of monocytes adherent to noninfected matrixes as well as monocytes adherent to S. sanguis-infected fibrin matrixes (Fig. 2B). Preincubation of rIL-10 with its neutralizing antibody suppressed the IL-10-mediated decrease in monocyte TFA (Fig. 2B). Monocytes adherent for 24 h to a noninfected or S. sanguis-infected fibrin matrix had TFA levels that were slightly lower than after 4 h of incubation (Fig.3). Addition of the anti-IL-10 antibody to the culture medium at the start of the culture period resulted in an increase in the TFA of monocytes incubated on noninfected fibrin and abrogation of the reduction of TFA of monocytes on S. sanguis-infected fibrin (Fig. 3).
TFA of monocytes incubated on noninfected or S. sanguis-infected fibrin matrixes for 4 (white bars) or 24 (shaded bars) h. Incubation for 24 h was either in the absence (black bars) or presence (gray bars) of 50 μg of anti-IL-10 MAb per ml to neutralize endogenously produced IL-10. Monocytes cultured for 4 h on noninfected fibrin matrixes had a TFA of 32.56 ± 11.26 pmol of FXa/1.5 × 106 cells. Results are means ± standard deviations of four experiments and are expressed as the fold increase in FXa formation compared to monocytes adherent to noninfected fibrin matrixes for 4 h (white bar). ∗, P = 0.05 , relative to noninfected fibrin matrix at 4 h, and P = 0.01 , relative to noninfected fibrin matrix at 24 h in the absence of anti-IL-10 MAb; ∗∗, P = 0.031 , relative to S. sanguis-infected fibrin matrix at 4 h, and P = 0.023 , relative to S. sanguis-infected matrix at 24 h in the absence of anti-IL-10 MAb.
IL-10 also inhibited the release of TNF-α, IL-1α, and IL-1β by monocytes incubated on S. sanguis-infected fibrin matrixes (Table 2).
Effect of rIL-10 on cytokine production by monocytes incubated on S. sanguis-infected fibrin matrixesa
Effect of CCS of monocytes on TFA and IL-10 production.Incubation of monocytes for 4 h on noninfected fibrin matrixes in the presence of CCS diluted in such a way that they contained 0.5 ng of TNF-α per ml, 21 pg of IL-1α per ml, and 93 pg of IL-1β per ml resulted in threefold higher levels of TFA (Fig.4) compared to monocytes incubated in medium. Supernatant obtained from a 4-h culture of S. sanguis on fibrin in the absence of monocytes did not induce monocyte TFA (data not shown). The role of TNF-α in CCS in the induction of monocyte TFA was assessed by incubation of CCS prior to use with a neutralizing anti-hTNF-α antibody. This neutralization of TNF-α did not reduce the ability of CCS to induce monocyte TFA (Fig.4). Incubation of monocytes with IL-1ra prior to addition of CCS, hampering the response to IL-1, also did not reduce monocyte TFA levels (Fig. 4). However, a combination of the two treatments led to a small but significant (P < 0.01) reduction of monocyte TFA (Fig. 4).
TFA of monocytes incubated for 4 h on a noninfected fibrin matrix in the presence or absence of CCS. The role of TNF-α and/or IL-1 in this CCS was determined by the use of anti-TNF-α MAb and/or IL-1ra as described in Materials and Methods. Monocytes cultured on noninfected fibrin matrixes in non-cytokine-containing medium had a TFA of 32.56 ± 11.26 pmol of FXa/1.5 × 106monocytes (white bar). Results are means ± standard deviations of three experiments and are expressed as the fold increase in FXa formation compared to monocytes adherent to noninfected fibrin matrixes (white bar). ∗, P < 0.01 , relative to normal non-cytokine-containing medium; ∗∗, P < 0.01 , relative to CCS in the absence of anti-TNF-α antibody and IL-1ra.
Incubation of monocytes for 24 h on noninfected fibrin matrixes with CCS resulted in a threefold increase in the IL-10 production (P < 0.001) compared to that of monocytes incubated in medium (Table 3). Preincubation of CCS with neutralizing anti-hTNF-α antibody reduced the IL-10 production by 43% ( P < 0.001; Table 3). Blocking the IL-1 receptor on monocytes by preincubation of these cells with IL-1ra had no effect on CCS-induced IL-10 production (Table 3). A combination of the two treatments did not decrease the IL-10 production any more than anti-hTNF-α treatment alone.
IL-10 production by monocytes incubated on noninfected fibrin matrixes with CCSa
DISCUSSION
In this study, we show that monocytes adherent to an S. sanguis-infected fibrin matrix express high levels of TFA which are accompanied by marked production of the proinflammatory cytokines TNF-α, IL-1α, and IL-1β and later, the anti-inflammatory cytokine IL-10. TFA as well as cytokine production was also found with monocytes that were adherent to noninfected fibrin; however, the levels of TFA and cytokine production were much lower.
The results of our study show that IL-10 plays an important role in dampening the TFA of monocytes adherent to an S. sanguis-infected fibrin matrix. This is most clearly demonstrated by the findings that rIL-10 reduces TFA and that neutralization of endogenously produced IL-10 by an anti-IL-10 monoclonal antibody (MAb) increases the TFA of monocytes cultured on an S. sanguis-infected fibrin matrix. These results are the first to demonstrate a role for IL-10 in TFA production in a relevant in vitro endocarditis model, thus extending reports by others who have shown that IL-10 diminishes LPS-induced (14, 18, 20, 29) or cytokine-induced (14, 24) monocyte and macrophage TFA. In our model, transforming growth factor beta, which has been reported by others (14, 18) to downregulate monocyte TFA, had no effect (Veltrop et al., unpublished data).
TNF-α and IL-1 are reported to induce monocyte TFA (17, 24, 25). However, the addition of either exogenous rTNF-α or rIL-1α to the incubation medium had no effect on monocyte TFA (during a 4-h culture period). This lack of effect was not due to biological inactivity of the cytokines, because similar concentrations of these cytokines were able to induce TFA on cultured human endothelial cells (37; unpublished data). Also, monocyte TFA was not diminished after neutralization of endogenous TNF-α produced during incubation on infected fibrin. Moreover, the experiments that show that the increased expression of TFA by monocytes incubated for 4 h with CCS was only marginally abrogated by the recombinant anti-TNF-α antibody and IL-1ra suggest that endogenous TNF-α and IL-1 may only play a minor role in the stimulation of TFA. Apparently, other, more potent stimuli are involved. These stimuli might be of bacterial origin, e.g., lipoteichoic acid, which is a potent activator of monocytes and human endothelial cells (10, 19, 21, 36). The present study, however, shows that incubation of fibrin-adherent monocytes with the supernatant obtained from a 4-h culture of fibrin-bound S. sanguis did not induce monocyte TFA (data not shown). These stimuli could also be of monocyte origin, as certain monocyte-derived factors, such as monocyte chemoattractant protein 1, IL-6, and IL-8, have been reported to induce monocyte TFA (23, 32). This awaits further investigation.
To further explore the mechanism by which IL-10 dampens the TFA induction, we first demonstrated that the production of TNF-α and IL-1 by fibrin-adherent monocytes was diminished by rIL-10. This is in line with the findings of others that IL-10 diminishes the LPS-mediated production of TNF-α and IL-1 by monocytes (11, 12, 15, 39). Thus, most probably the IL-10 produced by monocytes will, just like rIL-10, downregulate TNF-α and IL-1 production. These two cytokines, on the other hand, have been shown to induce IL-10 production by monocytes (15, 27, 28, 34, 40). It is therefore plausible to suggest that TNF-α and IL-1 produced initially by monocytes by their adherence to the fibrin matrix, and in particular to a fibrin matrix infected with S. sanguis, are responsible for the induction of IL-10 secretion somewhat later in time. Our findings show that CCS that contain the two proinflammatory cytokines, amongst other as yet undetermined factors, induced IL-10 production by fibrin-adherent monocytes (during a 24-h culture period). This production was in part dependent on TNF-α, since it was suppressed by more than 40% by neutralization of TNF-α in the supernatant. Using IL-1ra to inhibit the binding of IL-1 present in CCS to monocytes had no such effect. In contrast, the addition of rTNF-α or rIL-1α to the incubation medium did not induce monocyte IL-10 production. We cannot explain these observations presently. However, these data are in accordance with the results of Foey and coworkers (15), who also could induce IL-10 production by monocytes by the addition of endogenous TNF-α or IL-1 but not by the addition of rTNF-α or rIL-1.
The findings with CCS argue for a role of TNF-α in the IL-10 synthesis in human monocytes. This is in accordance with the findings of Wanidworanun and Strober (40). However, as we and others (28) have shown, TNF-α by itself is not sufficient to induce monocyte IL-10 production. Platzer and coworkers (28) could only detect IL-10 mRNA upon TNF-α stimulation of monocytes; for the translation of this mRNA a second stimulus was needed. This stimulus may be present in CCS, since the CCS induced IL-10 production by monocytes in a TNF-α-dependent manner.
In conclusion, we show that monocytes that adhere to a fibrin matrix, mimicking the in vivo vegetational surface, produce TFA, TNF-α, and IL-1 within 4 h. TNF-α and IL-1 play only a small, if any, role in the induction of monocyte TFA. The proinflammatory cytokine TNF-α plus other as yet unknown stimuli induce the monocytes to produce IL-10. In turn, IL-10 downregulates the production of TFA as well as that of TNF-α and IL-1. In this study we demonstrate that IL-10 is an important factor in the modulation of monocyte TFA on the surface of endocardial vegetations.
ACKNOWLEDGMENTS
This work was supported in part by The Netherlands Heart Foundation grant 91.058.
We thank H. Beekhuizen for reviewing the manuscript and G. P. Bieger-Smith for checking the manuscript for proper English usage.
Notes
Editor: E. I. Tuomanen
FOOTNOTES
- Received 20 September 2000.
- Returned for modification 21 December 2000.
- Accepted 5 February 2001.
- Copyright © 2001 American Society for Microbiology
