Lysophosphatidic Acid Inhibits CC Chemokine Ligand 5/RANTES Production by Blocking IRF-1–Mediated Gene Transcription in Human Bronchial Epithelial Cells

Lysophosphatidic acid (LPA) is a phospholipid mediator that exerts a variety of biological responses through specific G-protein–coupled receptors (LPA1–LPA5 and P2Y5). LPA is thought to be involved in airway inflammation by regulating the expression of anti-inflammatory and proinflammatory genes. Chemokines such as CCL5/RANTES are secreted from airway epithelium and play a key role in allergic airway inflammation. CCL5/RANTES is a chemoattractant for eosinophils, T lymphocytes, and monocytes and seems to exacerbate asthma. We stimulated CCL5/RANTES production in a human bronchial epithelial cell line, BEAS-2B, with IFN-γ and TNF-α. When LPA was added, CCL5/RANTES mRNA expression and protein secretion were inhibited, despite the presence of IFN-γ and TNF-α. The LPA effect was attenuated by Ki16425, a LPA1/LPA3 antagonist, but not by dioctylglycerol pyrophosphate 8:0, an LPA3 antagonist. Pertussis toxin, the inhibitors for PI3K and Akt also attenuated the inhibitory effect of LPA on CCL5/RANTES secretion. We also identify the transcription factor IFN regulatory factor-1 (IRF-1) as being essential for CCL5/RANTES production. Interestingly, LPA inhibited IFN-γ and TNF-α–induced IRF-1 activation by blocking the binding of IRF-1 to its DNA consensus sequence without changing IRF-1 induction and its nuclear translocation. Ki16425, pertussis toxin, and PI3K inhibitors attenuated the inhibitory effect of LPA on IRF-1 activation. Our results suggest that LPA inhibits IFN-γ– and TNF-α–induced CCL5/RANTES production in BEAS-2B cells by blocking the binding of IRF-1 to the CCL5/RANTES promoter. LPA1 coupled to Gi and activation of PI3K is required for this unique effect.

A sthma is a chronic inflammatory disease characterized by variable airflow, airway hyperresponsiveness, and accumulation of inflammatory cells to the airway (1). Eosinophils, lymphocytes, and mast cells participate in the process of complex airway inflammation that occurs in asthma. In particular, infiltration of eosinophils in bronchial mucosa is characteristic of most cases of asthma. Eosinophils produce a variety of toxic protein and lipid mediators such as cysteinyl-leukotrienes, which induce tissue injury and chronic airway inflammation. Accumulation of eosinophils to the airway partly depends on chemotaxis induced by chemokines. CCL5/RANTES is a member of the CC chemokine family and contributes to the pathogenesis of allergic diseases such as asthma, allergic rhinitis, and dermatitis (2)(3)(4). In asthma, the stimulated release of CCL5/RANTES from bronchial epithelia enables it to recruit T cells, macrophages, and eosinophils to the airway (5-7). CCL5/RANTES production from bronchial epithelial cells is induced synergistically by IFN-g and TNF-a, both of which play a pivotal role in chronic inflammation (3,8,9). IFN-g in bronchoalveolar lavage fluid (BALF) is actually increased in patients with asthma (10), and TNF-a is also induced in BALF after allergen challenge (11), indicating that IFN-g and TNF-a are most likely present at certain concentrations in the airway of asthmatics.
Lysophosphatidic acid (LPA) is a bioactive phospholipid mediator involved in cell proliferation and cell differentiation. LPA is generated from lysophosphatidylcholine by lysophospholipase D or autotaxin in the inflammation site (12). It is reported that LPA in BALF is increased in patients with asthma after allergen challenge (13) or pulmonary fibrosis (14). LPA stimulates the secretion of IL-8 from airway epithelium, inducing the infiltration of neutrophils (15). Furthermore, LPA has been reported to be involved in the production of angiogenic factors (16), pulmonary vascular injury (17), and the development of allergic airway inflammation (18). These findings suggest that LPA might accelerate inflammation in respiratory diseases such as acute lung injury/acute respiratory distress syndrome, pulmonary fibrosis, and asthma. In contrast, LPA also regulates the expression of anti-inflammatory genes in airway epithelium and may have a protective role in inflammation and remodeling (12). For example, LPA induces IL-13Ra 2 mRNA, protein expression, and secretion, which results in the attenuation of IL-13-mediated STAT6 phosphorylation, eotaxin gene expression, and suppressor of cytokine signaling 1 gene expression (19). LPA also induces cyclooxygenase-2 and PGE 2 production, which may function as an anti-inflammatory signal in the airway (20). Moreover, LPA has been shown to be involved in the enhancement of epithelial barrier function and the protection of lung injury (21). Thus, LPA seems to have dual roles in proinflammatory and anti-inflammatory signals that occur in respiratory systems. Having a similar proinflammatory and antiinflammatory role is sphingosine 1-phosphate (S1P). S1P is another bioactive lysophospholipid mediator that is also increased in BALF after allergen challenge and elicits contraction of airway smooth muscle cells (22). In contrast, S1P may protectively regulate eosinophilic inflammation and airway remodeling by inhibiting both CCL5/RANTES production and cell migration in human bronchial smooth muscle cells (23). We speculate that these lysophospholipid mediators have dual capacities against airway inflammation, acting as the defender or offender.
In the current study, we show that LPA significantly inhibits CCL5/RANTES production following costimulation with IFN-g and TNF-a in human bronchial epithelial cells. We explore this anti-inflammatory role of LPA and elucidate the intracellular mechanisms behind this unique inhibitory effect.

Quantitative real-time RT-PCR
Total RNA was extracted using the RNeasy Mini kit (Qiagen, Chatsworth, CA). TaqMan probes for real-time RT-PCR were purchased from Applied Biosystems (Foster City, CA). The RNA was treated with DNase, and then FIGURE 1. Effect of LPA on IFN-g-and TNF-a-stimulated CCL5/RANTES production in human bronchial epithelial cells. A, BEAS-2B cells were stimulated with the indicated concentration of IFN-g or TNF-a in the presence or absence of 1 mM LPA for 16 h. DMEM containing 0.1% BSA (0.1% BSA-DMEM) was used as control vehicle for IFN-g or TNF-a. CCL5/RANTES in the supernatant was measured by ELISA. The data are expressed as mean 6 SEM from nine independent experiments. B, BEAS-2B cells were stimulated with IFN-g (10 ng/ml) and TNF-a (10 ng/ml) in the presence of 0-10 mM LPA for 16 h. CCL5/RANTES in the supernatant was expressed as mean 6 SEM from six independent experiments. C, mRNA expression of CCL5/ RANTES in BEAS-2B cells was observed at 8 h postcostimulation with IFN-g and TNF-a. LPA (1 mM) significantly inhibited CCL5/RANTES mRNA expression compared with control vehicle (distilled water). CCL5/RANTES mRNA expression levels were standardized to GAPDH (mean 6 SEM, n = 6). D, NHBCs were stimulated for 16 h with or without IFN-g (10 ng/ml) and TNF-a (10 ng/ml) in the presence or absence of 1 mM LPA. CCL5/RANTES in the supernatant was expressed as mean 6 SEM from six independent experiments. pp , 0.05; ppp , 0.01. mRNA of human LPA 1 , LPA 2 , LPA 3 , LPA 4 , LPA 5 , P2Y5/LPA 6 , RANTES, IFN regulatory factor-1 (IRF-1), and GAPDH were measured by quantitative real-time RT-PCR using TaqMan gene expression assays on a sequence detection system model 7700 (Applied Biosystems). All data were standardized to GAPDH as previously described (23).

Immunoblotting
Posttreatment, the cells were lysed in buffer (10 mM Tris [pH 7.4], 150 mM NaCl, 1% Nonidet P-40, 0.1% sodium deoxycholate, 0.1% SDS, 1 mM EDTA, 10 mg/ml aprotinin, 5 mg/ml leupeptin, and 1 mM PMSF). Cell lysates were electrophoresed on 7.5% SDS-polyacrylamide gels and transferred to nitrocellulose membranes (Amersham Biosciences, Buckinghamshire, U.K.). The membranes were blocked in 1% skim milk overnight at 4˚C. Primary Abs for phospho-Akt (Ser 473 ) or total Akt (Cell Signaling Technology) in buffer containing 1% BSA were added and incubated overnight at 4˚C. The blots were washed three times for 5 min with PBS containing 0.1% Tween-20 and incubated for 1 h with anti-HRP Abs diluted in 5% skim milk. Blots were washed three times in PBS with 0.1% Tween-20 and detected by ECL plus regents (Amersham Pharmacia Biotech, Uppsala, Sweden). For immunoblotting IRF-1 in cell lysates, rabbit polyclonal IgG for IRF-1 (Santa Cruz Biotechnology, Santa Cruz, CA) was used as the primary Ab and rabbit polyclonal Ab specific for GAPDH (Santa Cruz Biotechnology) was used as the control. Nuclear proteins were extracted using a nuclear extraction kit (Panomics, Fremont, CA), and IRF-1 in nuclear extracts was detected using the method above.

Construction of adenoviruses and gene transfer
Recombinant adenovirus containing the regulator of G-protein signaling domain of p115Rho guanine nucleotide exchange factors (RhoGEFs) (p115RGS), C3 transferase, and enhanced GFP was constructed as described previously (23)(24)(25). Once BEAS-2B cells, in 12-or 24-well plates, were 50-70% confluent, the culture medium was changed to Opti-MEM (Invitrogen, San Diego, CA) containing adenoviruses. Cells were incubated for 1 h at 37˚C and then cultured for an additional 48 h in complete medium before being used in further experiments.

Transfection of small interfering RNA
BEAS-2B cells suspended in DMEM supplemented with 10% FBS were mixed with small interfering RNA (siRNA; Dharmacon, Lafayette, CO) and the RNAiMAX reagent (Invitrogen) and cultured in 6-or 12-well plates overnight. After this, the medium was replaced with complete medium, and cells were incubated for an additional 48 h before being used in further experiments.
Plasmids and luciferase assay IRF-1-and AP-1-luciferase reporter plasmids were purchased from Panomics. NF-kB-luciferase reporter plasmid was obtained from Stratagene (La Jolla, CA), and the NF-AT-luciferase reporter plasmid was from SABioscience (Frederick, MD). BEAS-2B cells were grown to 30-50% confluence in six-well plates and transfected with transcription factor-luciferase reporter plasmids using FuGENE 6 (Roche, Mannheim, Germany). The cells were incubated overnight in Opti-MEM and then cultured in complete medium for an additional 48 h. Luciferase activities in cell lysates were measured using the luciferase reporter assay system (Promega, Madison, WI). The consensus sequence of IRF-1 used in the IRF-1-luciferase reporter plasmid was: 59-GGAAGCGAAATGAAATTGACTGGAAGCGAAAATG-AAATTGACTGGAAGCGAAATGAATTGACTGGAAGCGAAAATGA-AATTGACTGGAAGCGAAAATGAAATTGACTGGAAGCGAAAATG-AAATTGACT-39.

EMSA
EMSA for IRF-1 was performed using a commercially available EMSA gelshift kit (Panomics) according to the manufacturer's instructions (26). Nuclear extracts (4 mg) were incubated with a biotinylated IRF-1-specific probe with or without the unlabeled probe for 30 min at 15˚C and then electrophoresed on a 6% polyacrylamide gel. Oligonucleotide-protein complexes were transferred to positively charged nylon membranes (Pall, Port Washington, NY). DNA was cross-linked to the membrane using a UV cross-linker (Stratagene) at 120 mJ/cm 2 , and bound DNA was detected with streptavidin-HRP and a chemiluminescence substrate. The sequence of the DNA probe specific for IRF-1 was: 59-GGAAGCGAAAATGAA-ATTGACT-39. To quantify the binding of IRF-1 to the DNA probe, the specific bands in the film were analyzed by a densitometer (Quantity One, Bio-Rad, Hercules, CA).

Statistical analysis
The data were analyzed using GraphPad Prism software (GraphPad, La Jolla, CA). A difference of the mean values between two independent groups was determined using Student t test. In the analysis of more than two groups, ANOVA was used to examine the statistical significance of differences, and when statistical significance was found, post hoc analysis (Bonferroni test) was performed. The p values , 0.05 were considered to be statistically significant.

Results
LPA inhibits CCL5/RANTES secretion from BEAS-2B cells following costimulation with IFN-g and TNF-a BEAS-2B cells were incubated for 8 h in 0.1% BSA-DMEM and stimulated with different concentrations of IFN-g and TNF-a. CCL5/RANTES secreted into the supernatant was measured at 16 h poststimulation. The higher levels of CCL5/RANTES secreted into the media were observed postcostimulation with 10 ng/ml IFN-g and 10 ng/ml TNF-a than from stimulation with IFN-g or TNF-a alone even at 100 ng/ml. When LPA or control vehicle was simultaneously added to the medium with IFN-g and/or TNF-a, 1 mM LPA was found to significantly inhibit CCL5/RANTES secretion (Fig. 1A). In addition, LPA inhibited CCL5/RANTES secretion in the presence of both IFN-g (10 ng/ml) and TNF-a (10 ng/ ml) in a dose-dependent manner. The maximal inhibitory effect of LPA on CCL5/RANTES secretion was attained at 1 mM (Fig. 1B). To examine mRNA expression, BEAS-2B cells were stimulated with 10 ng/ml IFN-g and 10 ng/ml TNF-a or control vehicle (0.1% BSA-DMEM) and incubated for an additional 8 h with or without 1 mM LPA. CCL5/RANTES mRNA expression was evaluated by real-time RT-PCR using TaqMan probes. We found that CCL5/ RANTES mRNA was markedly induced at 8 h postcostimulation with IFN-g and TNF-a. In contrast, LPA (1 mM) significantly inhibited the expression of CCL5/RANTES mRNA in BEAS-2B cells (Fig. 1C). The stimulatory effect of the combination of IFN-g and TNF-a and the inhibitory effect of LPA on CCL5/RANTES secretion are also observed in NHBCs (Fig. 1D).
The inhibitory effect of LPA on CCL5/RANTES secretion from BEAS-2B cells involves the activation of G i -protein and PI3K LPA receptors are coupled to heterotrimeric G-proteins. We investigated the LPA-mediated intracellular signaling pathway thought to be involved in inhibiting CCL5/RANTES secretion. PTX, which ADP ribosylates and inactivates G i -protein, significantly attenuated the inhibitory effect of LPA on CCL5/RANTES secretion (Fig. 3A). As G i -protein is linked to the sequential activation of PI3K, we investigated whether PI3K is involved in the inhibitory effect of LPA on CCL5/RANTES secretion. Both of the PI3K inhibitors, wortmannin (100 nM) and LY294002 (10 mM), significantly blocked the inhibitory effect of LPA on CCL5/ RANTES secretion (Fig. 3B). To ensure that LPA activated PI3K and Akt, we added LPA and then analyzed Ser 473 phosphorylation of Akt by immunoblotting using a phospho-Akt-specific Ab. Akt was phosphorylated at 5 min after the addition of 1 mM LPA in BEAS-2B cells (Fig. 3C). SH-5 (10 mM), a specific Akt inhibitor (28), also attenuated the inhibitory effect of LPA on CCL5/ RANTES secretion induced by costimulation with IFN-g and TNF-a in BEAS-2B cells (Fig. 3D).
Neither G 12/13 -, G q -, G s -protein nor the MEK/ERK pathway is involved in the inhibitory effects of LPA on CCL5/RANTES secretion from BEAS-2B cells We examined whether other types of G-proteins or MEK/ERK activation are involved in the inhibitory effect of LPA on CCL5/ RANTES secretion. Overexpression of p115RGS, which interacts with G 12 -and G 13 -protein, inactivates these G-proteins (25). C3 transferase inactivates a small G-protein, Rho, which is sequentially activated following G 12 /G 13 -protein activation. As shown in Fig. 4A, the inhibitory effect of LPA on IFN-g-and TNF-astimulated CCL5/RANTES secretion from BEAS-2B cells was not attenuated by either p115RGS or adenovirus-mediated overexpression of C3 transferase. However, both the overexpression of C3 transferase and p115RGS enhanced CCL5/RANTES secretion when stimulated with IFN-g and TNF-a. Furthermore, pretreatment with a G q -protein inhibitor, YM254890, had no effect on LPA inhibition (Fig. 4B). Forskolin, which activates adenylate cyclase, also did not affect the inhibitory effect of LPA on CCL5/RANTES secretion, suggesting that G s -protein signaling is not involved either (Fig. 4C). Pretreatment with MEK1/2 inhibitors PD98059 (30 mM) or U0126 (3 mM) significantly enhanced CCL5/RANTES secretion postcostimulation with IFN-g and TNF-a in BEAS-2B cells. However, MEK1/2 inhibitors did not attenuate the LPAinduced inhibition of CCL5/RANTES secretion (Fig. 4D).

LPA inhibits IRF-1-mediated gene transcription in BEAS-2B cells
We investigated activation of the transcription factors IRF-1, NF-kB, AP-1, and NF-AT, which bind to the promoter region of the CCL5/RANTES gene (29,30). First, we evaluated the activation of each transcription factor using luciferase reporter plasmids. These were designed to contain consensus DNA sequences of each transcription factor inserted upstream of the luciferase gene. Namely, the cis-acting enhancer element sequence resided upstream of the TATA box promoter that, upon transcription factor binding, drives expression of the firefly luciferase gene. BEAS-2B cells were stimulated with 10 ng/ml IFN-g and 10 ng/ ml TNF-a, and then incubated for 16 h with or without 1 mM LPA. Costimulation with IFN-g and TNF-a increased both IRF-1-and NF-kB-mediated luciferase activities. We found that LPA significantly inhibited IRF-1-mediated luciferase activity postcostimulation with IFN-g and TNF-a, yet without this stimulation, LPA alone did not affect IRF-1-mediated luciferase activity. In contrast, LPA induced NF-kB-mediated luciferase activity by itself and enhanced its activity postcostimulation with IFN-g and TNF-a. However, luciferase activities of the other two transcription factors, AP-1 and NF-AT, were hardly induced by costimulation with IFN-g and TNF-a. The effect of LPA on the luciferase activities of these transcription factors was also negligible (Fig. 5A). Because IRF-1-mediated luciferase activity was induced by costimulation with IFN-g and TNF-a, and then significantly inhibited by the addition of 1 mM LPA, we wanted to ascertain if IRF-1 was essential for CCL5/RANTES mRNA transcription. Knockdown experiments of IRF-1 mRNA and protein using IRF-1-specific siRNA were performed. We achieved knockdown of IRF-1 to ∼30% after siRNA treatment compared with levels from nonsilencing RNA (control siRNA) at 1 h postcostimulation with IFN-g and TNF-a (Fig. 5B). IRF-1 protein was also substantially decreased by siRNA treatment at 2 h postcostimulation with IFN-g and TNF-a (Fig. 5C). Reduction of IRF-1 by siRNA markedly attenuated CCL5/RANTES LPA inhibits IRF-1-mediated gene transcription via the LPA 1 receptor linked to the G i and PI3K-dependent intracellular signal transduction pathway As LPA 1 might be involved in the inhibitory effect of LPA on CCL5/RANTES production, we wanted to determine if antagonizing LPA 1 would attenuate the inhibitory effect of IRF-1 activation induced by costimulation with IFN-g and TNF-a. The LPA 1 /LPA 3 antagonist Ki16425, but not the LPA 3 antagonist DGPP 8:0, attenuated the inhibitory effect of LPA on IFN-g-and TNF-a-stimulated IRF-1 activation (Fig. 6A, 6B). We further investigated the roles of the G i -proteins and PI3K in the LPAmediated inhibition of IRF-1 activation. As expected, PTX as well as the PI3K inhibitors attenuated the inhibitory effect of LPA on IRF-1 activation costimulated with IFN-g and TNF-a (Fig. 6C, 6D).
LPA inhibits binding of IRF-1 to its consensus DNA sequence, but does not affect the induction of IRF-1 or translocation to the nucleus postcostimulation with IFN-g and TNF-a By immunoblotting, we demonstrated that costimulation with IFN-g and TNF-a induced IRF-1 in BEAS-2B cells. Induction of IRF-1 was obvious at 1-6 h poststimulation (Fig. 7A). LPA (1 mM) did not change the level of IRF-1 expression at 4 h postcostimulation (Fig.  7B). To determine whether LPA affected the translocation of IRF-1 to the nucleus or not, we performed immunoblotting on nuclear protein. We found that the amount of IRF-1 in the nucleus had increased at 4 h postcostimulation with IFN-g and TNF-a. However, LPA had no effect on the amount of IRF-1 in the nucleus 4 h poststimulation (Fig. 7C). By EMSA, we show that costimulation with IFN-g and TNF-a induces IRF-1 translocation to the nucleus and its sequential binding to its DNA consensus sequence. LPA (1 mM) decreased IRF-1 binding to the DNA, which was elicited by costimulation with IFN-g and TNF-a (Fig. 7D). We quantified the data from six independent EMSAs by densitometry and made sure B, IRF-1 mRNA expression at 1 h postcostimulation with IFN-g and TNF-a in BEAS-2B cells transfected with siRNA for IRF-1 was significantly decreased compared with its expression in cells transfected with control siRNA. mRNA expression for IRF-1 was assayed by quantitative real-time RT-PCR and standardized to GAPDH (means 6 SEM, n = 4). C, BEAS-2B cells transfected with either IRF-1 siRNA or control siRNA were stimulated with IFN-g and TNF-a. After 2 h incubation, total cell lysates were analyzed on a 10% SDS-polyacrylamide gel and immunoblotted with an anti-IRF-1 Ab. A representative photograph from six independent experiments is shown. D, Posttransfection of either control siRNA or IRF-1 siRNA, BEAS-2B cells were stimulated with IFN-g and TNF-a or control vehicle (0.1% BSA-DMEM) and incubated for 16 h. CCL5/RANTES in the supernatant was measured by ELISA. The data were standardized by the amount of CCL5/RANTES secreted from control siRNA-treated cells following costimulation with IFN-g and TNF-a (mean 6 SEM, n = 6). E, CCL5/RANTES mRNA expression was also examined at 8 h after the cells were stimulated with IFN-g and TNF-a or control vehicle (0.1% BSA-DMEM). The expression of CCL5/RANTES mRNA was assayed by quantitative real-time RT-PCR and standardized to GAPDH mRNA (mean 6 SEM, n = 4). pp , 0.05; ppp , 0.01. that LPA significantly decreased the binding of IRF-1 to the DNA sequence following costimulation with IFN-g and TNF-a (Fig. 7E). These results suggest that LPA does not affect either IRF-1 induction or translocation to the nucleus following costimulation with IFN-g and TNF-a, but does decrease the binding of IRF-1 to the DNA consensus sequence (Fig. 8).

Discussion
Airway epithelium is the first point of contact for inhaled allergens and microorganisms and plays a key role in the local immune response. It is also a considerable source of CCL5/RANTES, which accumulates inflammatory cells to the airway (31). Hence, CCL5/ RANTES production regulated by a variety of stimuli is involved in the pathology of airway diseases such as asthma (2)(3)(4). LPA, a bioactive lipid molecule, is elevated in BALF of patients with respiratory diseases and is suggested to be involved in the regulation of airway functions by acting as a proinflammatory and anti-inflammatory mediator (12). In the current study, we show for the first time, to our knowledge, that LPA inhibits CCL5/RANTES production in response to IFN-g and TNF-a stimulation in a human bronchial epithelial cell. Our results indicate that LPA-induced inhibition of the expression of CCL5/RANTES mRNA and protein is associated with the inhibition of IRF-1 transcriptional activity via the LPA 1 receptor/G i -proteins and mechanisms involving PI3K and Akt (Fig. 8). This inhibitory effect on CCL5/ RANTES production may reflect the anti-inflammatory role of LPA in the respiratory system.
The mode of action for extracellular LPA is the activation of specific seven-transmembrane domain G-protein-coupled receptors. In BEAS-2B cells and NHBCs, we were able to detect the mRNA of LPA 1 , LPA 2 , LPA 3 , LPA 5 , and P2Y5/LPA 6 but not LPA 4 . The LPA 1 /LPA 3 antagonist Ki16425 but not the LPA 3 antagonist DGPP 8:0 attenuated the inhibitory effect of LPA on CCL5/RANTES production. This suggests that this unique effect of LPA is mediated by LPA 1 . It is reported that LPA 1 links with G 12/13 -, G i -, and G q -proteins (32). We show that the inhibitory effect of LPA on CCL5/RANTES production is mediated by G i -protein activation linked to the PI3K-Akt-dependent intracellular signaling pathway. This was demonstrated using PTX and inhibitors for PI3K and Akt to negate the LPA effect. It is known that either activation of G i or PI3K is involved in the inhibitory effects on IL-12 family cytokine production from activated monocytes and macrophages in a similar way to our system. Members of the MCP-1 chemokine family suppress IL-12 production from primary human monocytes through CCR2, and this suppression is partially recovered by PTX (33,34). C5a negatively regulates TLR4induced IL-12 and IL-23 production in murine macrophages through a PI3K-dependent pathway (35). Adenylate cyclase activator forskolin and G q inhibitor YM254890 failed to attenuate the inhibitory effect of LPA on CCL5/RANTES production. This suggests that neither G s nor G q activation is involved in LPAmediated inhibition. Even though overexpression of C3 transferase and the RGS domain of p115RhoGEF enhanced CCL5/ RANTES production, LPA was still effective in inhibiting it. This suggests that the activation of G 12 /G 13 -proteins and the Rho signaling pathway negatively regulates CCL5/RANTES production. However, it seems that neither G 12 /G 13 -proteins nor Rho is involved in the LPA-induced inhibitory effect on CCL5/RANTES production. Consistent with a previous study (36), MEK1/2 inhibitors enhanced CCL5/RANTES production following costimulation with IFN-g and TNF-a. However, MEK1/2 inhibitors did not attenuate the inhibitory effect of LPA on CCL5/RANTES production. Therefore, the MEK/ERK pathway seems to regulate CCL5/RANTES production negatively independent of the LPAinduced PI3K-dependent pathway, which also negatively regulates CCL5/RANTES production.
IRF-1 is induced by a number of stimuli, including IFN-g, TNFa, and LPS, in a variety of cells (37)(38)(39) and is supposed to be an essential transcriptional activator for IFN-g-induced CCL5/ RANTES expression in murine macrophages (40,41). Consistent with previous reports (9,42), costimulation with IFN-g and TNF-a induced CCL5/RANTES production in BEAS-2B cells and was associated with IRF-1 expression (Fig. 7A). This suggests that IRF-1 may also regulate CCL5/RANTES gene transcription FIGURE 6. Effects of LPA receptor antagonists, PI3K inhibitors, and PTX on LPA-induced inactivation of IRF-1. The IRF-1 luciferase reporter plasmid was transiently transfected into BEAS-2B cells. The transfected cells were used for further experiments. The cells were treated with 1 mM Ki16425 or control vehicle (DMSO) for 30 min in A, with 1 mM DGPP 8:0 or control vehicle (distilled water) for 30 min in B, with 100 ng/ml PTX or control vehicle (distilled water supplemented with 0.1% BSA) overnight in C, and with 100 nM wortmannin, 10 mM LY294002, or control vehicle (DMSO) for 30 min in D. The cells were then stimulated for 16 h with 10 ng/ml IFN-g and 10 ng/ml TNF-a in the presence or absence of 1 mM LPA. Luciferase activities in cell lysates were assayed and standardized against cells stimulated with IFN-g and TNF-a, without LPA (mean 6 SEM, n = 6). ppp , 0.01. finding that LPA does not attenuate the expression level of IRF-1 protein induced by IFN-g and TNF-a. The nuclear content of IRF-1 was not altered by LPA, indicating that IRF-1 binding to DNA rather than nuclear transport of IRF-1 is attenuated by the LPA signaling pathway.
Although the PI3K-Akt-dependent pathway seems to be involved in the regulation of CCL5/RANTES gene expression, the mechanisms by which this intracellular signaling pathway affects the interaction of IRF-1 with the DNA or its transcriptional activity remain unknown. We speculate that LPA affects the interaction of IRF-1 with other factors that regulate DNA binding, thereby inhibiting its transcriptional activity. For example, the phosphorylation status of IRF-1 or the association of IRF-1 with coactivators and IRF-1 binding proteins might be affected by LPA. In relation to this, it is reported that IRF-1 is phosphorylated by casein kinase II, a serine/threonine protein kinase, which is essential for IRF-1 binding and NF-kB synergistic activation on the IFN-b promoter (44,45). Further experiments are needed to clarify the mechanism behind the inhibition of IRF-1 binding to DNA that is mediated by LPA signaling.
In conclusion, LPA has a unique inhibitory effect on IRF-1 transcriptional activity following costimulation with IFN-g and TNF-a. This effect may elicit the decrease of CCL5/RANTES secretion in human airways, thereby modulating chronic inflammation that is observed in patients with bronchial asthma.