LPS has a priming effect on various stimuli. For instance, LPS priming enhances the production of platelet-activating factor (PAF), a proinflammatory lipid mediator that is induced by PAF itself. Among various enzymes responsible for PAF biosynthesis, acetyl-coenzyme A:1-O-alkyl-2-lyso-sn-glycero-3-phosphocholine acetyltransferase is one of the enzymes activated by PAF receptor stimulation. In this study we investigated the priming effect of LPS on the acetyltransferase activation by PAF in TLR4-knockout (KO) mice, MyD88-KO mice, and Toll/IL-1R domain-containing adaptor inducing IFN-β (TRIF)-KO mice. This enzyme was biphasically activated by LPS. Although the first peak occurred within 30 min in wild-type (WT), but not TLR4-KO or MyD88-KO, macrophages, the second phase reached a maximum within hours in WT, MyD88-KO, and TRIF-KO, but not in TLR4-KO, macrophages. Only in the second phase was the increase in acetyltransferase activity upon PAF receptor activation remarkably enhanced in WT, MyD88-KO, and TRIF-KO cells, but not in TLR4-KO cells. These data demonstrated that LPS exerted a priming effect on PAF receptor-mediated acetyltransferase activation through the TLR4-dependent, but MyD88- and TRIF-independent, pathway.
Platelet-activating factor (PAF),3 1-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine, is a potent lipid mediator eliciting responses in a wide variety of cell types (1, 2, 3). PAF is involved in numerous biological responses, including platelet activation, airway constriction, and hypotension (1, 2, 3). The biosynthesis of PAF has been extensively studied in various cells and tissues (1, 3). PAF is synthesized via two distinct pathways: the de novo and remodeling pathways (3, 4, 5). The latter is considered to be regulated by extracellular signals, and it plays a critical role in stimulus-coupling PAF biosynthesis (6, 7, 8, 9, 10, 11, 12). PAF synthesis induced by extracellular signals has been reported in mouse peritoneal cells stimulated by calcium ionophore (12), rat peritoneal cells stimulated by PAF (9), human eosinophils stimulated by fMLP (7), human mesangial cells stimulated by LPS (10), and IC-21 mouse peritoneal macrophage cells stimulated by LPS (11). In the remodeling pathway, the precursor of PAF, 1-O-alkyl-sn-glycero-3-phosphocholine (lyso-PAF), is synthesized from 1-O-alkyl-2-arachidonoyl-sn-glycero-3-phosphocholine and/or 1-O-alk-1-enyl-2-arachidonoyl-sn-glycero-3-phosphoethanolamine (ethanolamine plasmalogen) (4, 5, 13) by the action of phospholipase A2 (PLA2). Subsequently, lyso-PAF is acetylated by acetyl-coenzyme A (CoA):lyso-PAF acetyltransferase (lyso-PAF acetyltransferase) to form PAF (4, 5, 14).
Lyso-PAF acetyltransferase was initially demonstrated and partially characterized by Wykle et al. (14). Phosphorylation of the enzyme was reported to modulate the activity of human neutrophils (15) and rat splenic microsomes (16, 17), although the precise mechanisms of the enzyme activation are poorly understood. Although lyso-PAF acetyltransferase has been detected in microsomes of rat spleen, lung, thymus, and porcine leukocytes (14, 18, 19), it has not been purified or cDNA cloned. Recently, lyso-PAF acetyltransferase of human neutrophils was reported to be activated by either TNF-α or fMLP stimulation in the p38 MAPK-dependent pathway (20). In addition, activation of acetyltransferase by LPS was also shown in IC-21 macrophage cells (11).
LPS activates TLR4, one of the TLRs that play a central role in activation of the innate immune system (21, 22, 23, 24, 25). Through a Toll/IL-1R (TIR) domain-containing adaptor molecule named MyD88, the TLR4 signaling pathway leads to the induction of cyclooxygenase-2 (26) and inflammatory cytokines such as IL-12 and IL-6 (22, 23, 24, 26). TLR4 also regulates the activation of IFN-regulatory factor 3 and induction of IFN-β via the other adaptor molecules, i.e., TIR domain-containing adaptor inducing IFN-β (TRIF) and TRIF-related adaptor molecule, in the MyD88-independent pathway (22, 23, 24, 25, 26, 27, 28). Although NF-κB transcription factor and MAPKs were activated by LPS in both pathways, signaling via the MyD88-independent pathway was delayed compared with that via the MyD88-dependent pathway (26, 29).
Worthen et al. (30) previously demonstrated that treatment of human neutrophils with LPS resulted in enhanced production of PAF upon stimulation with fMLP. This LPS-enhanced effect is called priming. Several similar phenomena have been observed. LPS primed the neutrophils such that they responded highly to subsequent fMLP stimulation with enhanced secretion of superoxide anion, O⨪2 (31), and lysosomal enzymes (32). After stimulation with C5a or fMLP, rabbit neutrophils pretreated with LPS were more injurious to cultured endothelial cells than those without LPS treatment (33). LPS priming in P388D1 mouse macrophage cells enhanced arachidonic acid release and PGE2 production by stimulants such as PAF (34). Moreover, Heuer et al. (35) showed that pretreatment of mice with LPS enhanced sensitivity to PAF, resulting in increased mortality. However, the mechanisms as well as the molecules involved in the priming effect of LPS remain quite obscure.
Materials and Methods
Methylcarbamyl PAF (mc-PAF), 1-O-hexadecyl-2-N-methylcarbamyl-sn-glycero-3-phosphocholine, and lyso-PAF (1-O-hexadecyl-sn-glycero-3-phosphocholine) were purchased from Cayman Chemical. LPS from Salmonella minnesota and 4-amidinophenylmethanesulfonyl fluoride (APMSF) were purchased from Sigma-Aldrich. [3H]Acetyl-CoA (185 GBq/mmol) was obtained from Amersham Biosciences. SB 203580 and SP 600125 were purchased from Tocris Cookson. SB 202190 and SB 202474 were obtained from Calbiochem. U0126 was purchased from Cell Signaling Technology. These MAPK inhibitors were dissolved in DMSO. Anti-p38 or anti-phospho p38 MAPK Ab were obtained from Cell Signaling Technology. Recombinant mouse IFN-β (Silkworm larvae) was purchased from Katakura Industries. Anti-mouse IFN-β Ab (clone 7F-D3, a rat IgG1) was obtained from Yamasa. A proteinase inhibitor mixture, Complete, was purchased from Roche. Cycloheximide was purchased from Wako Pure Chemical Industries. Bondapak C8 was obtained from Millipore.
C57BL/6J mice were obtained from Clea Japan. TLR4-KO (21) and MyD88-KO (36) mice were obtained by backcrossing onto C57BL/6J mice seven and eight times, respectively. TRIF-KO mice (27) were maintained on a mixed C57BL/6 × 129/Ola genetic background. Male and female mice, 9–24 wk of age, were used. In each experiment analyzing TLR4-KO, MyD88-KO, and TRIF-KO mice, sex- and age-matched wild-type (WT) mice were used as controls. Mice were maintained in a light-dark cycle with lights on from 0700–2000 h at 22°C. Mice were fed a standard laboratory diet and water ad libitum. All animal studies were conducted in accordance with the guidelines for animal research at University of Tokyo and were approved by the University of Tokyo ethics committee for animal experiments.
Isolation and stimulation of mouse peritoneal macrophages
Three days after i.p. injection of 2 ml of 4% thioglycolate, peritoneal exudate cells were harvested from the peritoneal cavity three times with 5 ml each of ice-cold PBS. The cells were washed twice with 30 ml of ice-cold PBS after centrifugation at 250 × g at 4°C. Total cell numbers were determined with a hemocytometer after staining with Turk solution (Wako Pure Chemical Industries). Cells (3.2 × 106 cells/6-cm dish) were cultured in RPMI 1640 medium supplemented with 10% FBS (RPMI 1640–10% FBS) at 37°C in 5% CO2. After incubation for 2 h, the medium was changed and additionally cultured in 4 ml of RPMI 1640–10% FBS for 20–22 h. The cells were stained with Diff-Quick (International Reagents), and morphological observation by microscopy revealed that most adherent cells were macrophages. The cells were treated with 100 ng/ml LPS or 200 nM mc-PAF (methylcarbamyl PAF, a nonhydrolyzable PAF analog) for the indicated periods. After that treatment, they were washed with an ice-cold buffer containing 20 mM Tris-HCl (pH 7.4) and 0.3 M sucrose. Subsequently, the buffer was discarded, and the cells were frozen with liquid nitrogen. For preparation of cell extracts, cells were scraped into 600 μl of an ice-cold buffer containing 20 mM Tris-HCl, (pH 7.4), 50 mM β-glycerophosphate, 1 mM sodium orthovanadate, 5 mM 2-ME, 20 μM APMSF, and a proteinase inhibitor mixture, and the collected cells were sonicated twice on ice for 30 s each time. Intact cells, cellular debris, and mitochondria were removed by centrifugation at 9000 × g for 10 min at 4°C. The protein concentration of the supernatant was measured by the method of Bradford (37), using protein assay solution (Bio-Rad) and BSA (fraction V, fatty acid-free; Sigma-Aldrich) as a standard.
TRIF-KO cells were stimulated with LPS for 4 h. After stimulation, total RNA was prepared using the Absolutely RNA RT-PCR Miniprep kit (Stratagene). Total RNA of macrophages was used as a template to amplify IFN-γ-inducible protein 10 mRNA/by IFN through the TRIF-dependent pathway, by OneStep RT-PCR (Qiagen).
Effects of MAPK inhibitors and cycloheximide
Macrophages were pretreated with the indicated concentrations of MAPK inhibitors for 1 h and subsequently stimulated with 100 ng/ml LPS for 30 min. For the cycloheximide experiments, macrophages were pretreated with 10 μg/ml cycloheximide for 1 h and subsequently stimulated with 100 ng/ml LPS for 2 h. The medium was changed and additionally cultured with LPS for 14 h in the absence of cycloheximide, because cycloheximide was toxic to macrophages.
Western blot analysis
Total protein (10.6 μg) in the cell extracts was resolved by SDS-PAGE with 10% polyacrylamide separating gels and transferred to a Hybond ECL nitrocellulose membrane (Amersham Biosciences). The membrane was blocked with Block Ace (Dainippon Pharmaceuticals), probed with anti-p38 or anti-phospho p38 MAPK rabbit polyclonal Ab, washed, and incubated with HRP-linked anti-rabbit IgG. After washing, the immunoreactive proteins were detected using ECL reagents (Amersham Biosciences) and x-ray film (Amersham Biosciences).
Assay of lyso-PAF acetyltransferase
The assay was modified from the method described by Kume et al. (19). The reaction mixture contained 20 mM Tris-HCl (pH 7.4), 2 mM CaCl2, 1 mg/ml phosphatidylcholine (Sigma-Aldrich), 100 μM [3H]acetyl-CoA (1.11 GBq/mmol), 5 mM 2-ME, 20 μM APMSF, a proteinase inhibitor mixture, and 18 μg of proteins with or without 20 μM lyso-PAF in a total volume of 100 μl. The enzyme reactions were performed in a water bath at 37°C. After a 7.5-min incubation, 122 μl of ice-cold methanol was added to terminate the reaction. The products were bound to 6 mg of Bondapak C8 resin, which were washed eight times with 200 μl of 55% methanol in 20 mM Tris-HCl (pH 7.4). The radioactive products were eluted with 250 μl of methanol. After drying up at 50°C for 3 h, the radioactivities were measured with a TopCount scintillation counter (PerkinElmer) in the presence of 200 μl of Microscinti-0 (PerkinElmer). The difference between the radioactivities obtained with and without lyso-PAF corresponded to lyso-PAF acetyltransferase activity.
Priming effect on lyso-PAF acetyltransferase activity
Peritoneal macrophages were pretreated with 100 ng/ml LPS or 100 U/ml IFN-β for the indicated periods and subsequently stimulated with 200 nM mc-PAF for 1 min. The anti-IFN-β Ab (1/500) was added 30 min before LPS treatment.
Values were expressed as the mean ± SEM. A value of p < 0.05 was taken to be statistically significant. All statistical calculations were performed with StatView-J, version 5.0 (Abacus Concepts).
Activation of lyso-PAF acetyltransferase in macrophages after LPS stimulation
Peritoneal macrophages from WT, TLR4-KO, and MyD88-KO mice were stimulated with LPS from 0 to 120 min. The activity of acetyltransferase reached a peak at 30 min in WT cells and declined as the incubation continued until 120 min. Neither TLR4-KO nor MyD88-KO cells were responsive to LPS (Fig. 1, A and B). In TRIF-KO cells, however, acetyltransferase was activated 30 min after LPS stimulation at a similar level to WT cells (Fig. 1 C). TRIF-KO cells were practically defective in TRIF-dependent responses to LPS, as judged by the severe reduction of IFN-γ-inducible gene mRNA induction by LPS (data not shown), consistent with previous data (27).
Effects of MAPK inhibitors on the activation of lyso-PAF acetyltransferase
It is known that MAPKs are phosphorylated by LPS stimulation (22, 26, 29). To identify MAPKs involved in the acetyltransferase activation by LPS, we used several specific inhibitors of p38 (SB 203580 and SB 202190), JNK (SP 600125), and MEK1/2 (U0126; Fig. 2). Macrophages were preincubated in the presence or the absence of MAPK inhibitors for 1 h and subsequently stimulated with LPS for 30 min. The LPS-induced acetyltransferase activation was almost abolished in the presence of p38 inhibitors, SB 203580 and SB 202190, but was unaffected by the inactive analog SB 202474. Consistently, when cell lysates were analyzed by Western blot using an Ab against the phospho- or total form of the p38 MAPK, the amount of phospho-p38 MAPK reached a peak at 30 min after LPS stimulation (Fig. 3). Neither SP 600125 nor U0126 abolished LPS-induced enzyme activation (Fig. 2). LPS-induced phosphorylation of c-Jun and ERK1/2, which are substrates of JNK and MEK1/2, were abolished in the presence of SP600125 and U0126, respectively (data not shown).
Activation of lyso-PAF acetyltransferase in macrophages after PAF receptor stimulation
Lyso-PAF acetyltransferase was previously reported to be activated by PAF stimulation in rat peritoneal neutrophils (9). In peritoneal macrophages from WT, TLR4-KO, and MyD88-KO mice, we examined the elevation of acetyltransferase activity at 0–240 s after PAF receptor stimulation (Fig. 4). In this study we used mc-PAF, a nonhydrolyzable PAF analog, because PAF is degraded by PAF acetylhydrolases in serum (38, 39). In all genotypes, the enzyme was rapidly activated, with a maximum at 30 s after mc-PAF stimulation. The mc-PAF-induced enzyme activation was not suppressed by any of the inhibitors of p38, JNK, or MEK1/2 (data not shown).
Priming effect of LPS on mc-PAF-induced acetyltransferase activity
WT macrophages were treated with LPS for 1, 4, 8, and 16 h. At each time point, the cells were stimulated with mc-PAF for 1 min (Fig. 5,A). This procedure reveals the priming effect of LPS on PAF receptor-mediated acetyltransferase activation. We found that such a lengthy (8 and 16 h) treatment of cells with LPS alone elicited a second-phase increase in the enzyme activity. Even though the response to mc-PAF alone remained constant during the assay, LPS markedly potentiated the enzyme activation by mc-PAF at 8 and 16 h. At 16 h, acetyltransferase activities were increased 4.1 ± 0.7-, 2.8 ± 0.3-, and 9.7 ± 1.5-fold (mean ± SEM; n = 6) in LPS alone, mc-PAF alone, and both stimulations, respectively, compared with nonstimulated cells. Although the enzyme was biphasically activated at 30 min and 16 h after LPS treatment (Fig. 1 and 5, A and B), the priming effect of LPS on PAF was only observed at 16 h and was absent at 30 min (Fig. 5,B). The priming effect of LPS as well as the LPS-induced acetyltransferase activation were almost abolished by cycloheximide (Fig. 5 C).
Priming effect of LPS in the MyD88- and TRIF-independent pathway
Unlike WT macrophages, LPS treatment did not enhance PAF receptor-mediated acetyltransferase activation in TLR4-KO macrophages (Fig. 6,A). However, LPS-treated MyD88-KO cells and TRIF-KO cells displayed a marked increase in the enzyme activity by mc-PAF, at a similar level to WT cells (Fig. 6, B and C). A low dose of LPS (1 ng/ml) also displayed the priming effect in both MyD88-KO and TRIF-KO cells at a similar level to WT cells (data not shown).
In this report we indicated that lyso-PAF acetyltransferase was activated 1) by a second-order time course after PAF receptor stimulation, 2) by a minute-order time course after LPS stimulation in the MyD88- and p38 MAPK-dependent pathway, and 3) by an hour-order time course after LPS-stimulation in the MyD88- and TRIF-independent pathway (Figs. 6, B and C, and 7A). Furthermore, this study demonstrates for the first time that the LPS-priming effect on acetyltransferase activation occurs through the TLR4-dependent, but MyD88- and TRIF-independent, pathway (Fig. 7 B).
Although the activity of lyso-PAF acetyltransferase reached a peak at 30 min with LPS treatment in WT cells, neither TLR4-KO nor MyD88-KO cells displayed the enzyme activation by LPS (Fig. 1, A and B). These results indicate that the TLR4/MyD88-dependent pathway is responsible for LPS-induced acetyltransferase activation. TLR4 binds MyD88 via TIR domain-containing adaptor protein (28). MyD88 possesses the death domain, which mediates the association with IL-1R-associated kinase, which is then phosphorylated. The phosphorylated IL-1R-associated kinase associates with TNFR-activated factor 6 and activates the MAPKs. This process leads to the activation of p38 MAPK (22, 23, 24). We showed in this study that LPS-induced acetyltransferase activation was almost abolished in the presence of p38 MAPK inhibitors (Fig. 2). In addition, phosphorylation of p38 MAPK peaked at 30 min after LPS stimulation (Fig. 3) in parallel with acetyltransferase activity (Fig. 1). These results demonstrate that lyso-PAF acetyltransferase of macrophages is activated by LPS stimulation in the p38 MAPK-dependent pathway. In MyD88-KO mice, although production of cytokines, IL-1β, TNF-α, and IL-6 in response to LPS was completely impaired, activation of p38 was delayed (from 10 to 20 min) but not attenuated, compared with WT (26, 29). However, our study showed that lyso-PAF acetyltransferase of MyD88-KO macrophages was not activated, at least within 120 min after LPS stimulation (Fig. 1 B). In human macrophages, expressions of p38α and -δ mRNA were high, but those of β and γ mRNA were low (40). Additionally, SB 203580 and SB 202190 inhibited the enzymatic activity of p38α and -β isoforms, but did not for γ or δ (41, 42). Thus, our data suggest that lyso-PAF acetyltransferase is activated mainly by the p38α isoform. p38α is unlikely to be activated by the MyD88-independent pathway.
By mc-PAF stimulation, lyso-PAF acetyltransferase was rapidly activated in WT, TLR4-KO, and MyD88-KO macrophages, with a maximum at 30 s (Fig. 4). These results suggest that the enzyme is normally expressed and activated by PAF in TLR4- and MyD88-KO mice. The peak of acetyltransferase activation by mc-PAF stimulation occurred much earlier than that by LPS stimulation. Furthermore, mc-PAF-induced acetyltransferase activation was not suppressed by any of the inhibitors of p38, JNK, and MEK1/2. PAF appears to activate lyso-PAF acetyltransferase in a MAPK-independent, and possibly Ca2+-dependent, manner (43, 44). However, cellular signaling to acetyltransferase activation by PAF remains to be clarified.
To investigate the priming effect of LPS, WT macrophages were treated with LPS for >8 h and were subsequently stimulated with mc-PAF (Fig. 5,A). We found that such a long treatment with LPS alone elicited the second phase of acetyltransferase activation. Furthermore, LPS markedly potentiated the acetyltransferase activation by mc-PAF at 8 and 16 h. This kind of biological response has been called LPS priming (30, 31, 32, 33, 34, 35). Although lyso-PAF acetyltransferase was biphasically activated 30 min and 16 h after LPS treatment (Fig. 1, 5, A and B), the priming effect of LPS on PAF was only observed at the second phase (Fig. 5,B). It is noteworthy that both MyD88-KO cells and TRIF-KO cells exhibited such a priming effect; LPS-treated MyD88-KO cells and TRIF-KO cells displayed marked increases in the acetyltransferase activity produced by mc-PAF at levels comparable to WT cells (Fig. 6, B and C). These results showed that MyD88 and TRIF were dispensable for LPS priming of mc-PAF-induced acetyltransferase activation. MyD88 is a critical adaptor protein for TLR4 signaling (22, 23, 24, 25). TLR4 uses TRIF as another adaptor molecule to induce the expression of IFN-β in a MyD88-independent manner (24, 27). After the binding of IFN-β to its specific cell surface receptor (IFN-αβR), a series of IFN-inducible genes are activated (24, 45). Consistent with the data for TRIF-KO macrophages, IFN-β could not replace LPS in priming WT macrophages, and anti-IFN-β Ab did not suppress the LPS-priming effect on WT cells (data not shown). It has been reported that the LPS-TLR4 pathway involves either the MyD88- or the TRIF/TRIF-related adaptor molecule-IFN-β pathway (24, 27, 28). Our pharmacological and genetic data strongly suggest that there may be an alternative, as yet unidentified, pathway(s) in LPS priming of lyso-PAF acetyltransferase activation. It is clear that de novo protein synthesis is required for the response, because the priming effect was inhibited by cycloheximide treatment of the cells (Fig. 5 C). Additional studies are required for elucidation of the pathways and molecules downstream of TLR4.
Mariano et al. (46) showed that blood and urinary concentrations of PAF were significantly higher in septic patients than in control subjects. In humans administered LPS, treatment with the PAF receptor antagonist Ro 24–4736 resulted in fewer symptoms, such as rigor and myalgia, than those with a placebo (47). These reports suggest the clinical significance of PAF in sepsis. We recently demonstrated the importance of cytosolic PLA2α (cPLA2α) in a mouse model of adult respiratory distress syndrome caused by sepsis, using both cPLA2α-deficient mice (48) and a cPLA2α inhibitor (49). Because cPLA2α is also a crucial enzyme for PAF synthesis via the remodeling pathway (12, 50), it is possible that PAF is involved in the pathology of sepsis-induced adult respiratory distress syndrome. Therefore, many symptoms of sepsis may be ameliorated by restricting the MyD88- and TRIF-independent pathway that leads to PAF production by LPS priming.
We are grateful to Drs. T. Nagase, T. Yokomizo, N. Uozumi, and H. Nishina (University of Tokyo) for valuable suggestions, and to K. Noguchi, C. Ohkawara, T. Sato, Y. Takahashi, Y. Yamada, Y. Kitaike, and F. Hamano (University of Tokyo) for technical assistance. We also thank John H. Jennings for critical reading of the manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported in part by grants-in-aid from the Ministry of Education, Science, Culture, Sports, and Technology of Japan (to T.S. and S.I.); Grants-in-Aid for Comprehensive Research on Aging and Health and for Research on Allergic Disease and Immunology from the Ministry of Health, Labor, and Welfare of Japan; and grants from the Yamanouchi Foundation for Research on Metabolic Disorders, the Kanae Foundation for Life and Socio-Medical Science, the Uehara Memorial Foundation (to S.I.), and the Welfide Medicinal Research Foundation (to H.S.). H.S. is a research fellow of the Japan Society for the Promotion of Science.
Abbreviations used in this paper: PAF, platelet-activating factor; APMSF, 4-amidinophenylmethanesulfonyl fluoride; CoA, coenzyme A; cPLA2α, cytosolic PLA2α; lyso-PAF, 1-O-alkyl-sn-glycero-3-phosphocholine; MEK, MAPK kinase; PLA2, phospholipase A2; mc-PAF, methylcarbamyl PAF; TIR, Toll/IL-1 receptor; TRIF, TIR domain-containing adaptor inducing IFN-β; WT, wild type.