Phospholipase D Activation by Sphingosine 1-Phosphate Regulates Interleukin-8 Secretion in Human Bronchial Epithelial Cells*

Sphingosine 1-phosphate (S1P), a potent bioactive sphingolipid, has been implicated in many critical cellular events, including a regulatory role in the pathogenesis of airway inflammation. We investigated the participation of S1P as an inflammatory mediator by assessing interleukin-8 (IL-8) secretion and phospholipase D (PLD) activation in human bronchial epithelial cells (Beas-2B). S1P1, S1P3, S1P4, S1P5, and weak S1P2 receptors were detected in Beas-2B and primary human bronchial epithelial cells. S1P stimulated a rapid activation of PLD, which was nearly abolished by pertussis toxin (PTX) treatment, consistent with S1P receptor/Gi protein coupling. S1P also markedly induced Beas-2B secretion of IL-8, a powerful neutrophil chemoattractant and activator, in a PTX-sensitive manner. This S1P-mediated response was dependent on transcription as indicated by a strong induction of IL-8 promoter-mediated luciferase activity in transfected Beas-2B cells and a complete inhibition by actinomycin D. Beas-2B exposure to 1-butanol, which converts the PLD-generated phosphatidic acid (PA) to phosphatidylbutanol by a transphosphatidylation reaction, significantly attenuated the S1P-induced IL-8 secretion, indicating the involvement of PLD-derived PA in the signaling pathway. Inhibition of 12-O-tetradecanoyl-phorbol-13-acetate-stimulated IL-8 production by 1-butanol further strengthened this observation. Blocking protein kinase C and Rho kinase also attenuated S1P-induced IL-8 secretion. Our data suggest that PLD-derived PA, protein kinase C, and Rho are important signaling components in S1P-mediated IL-8 secretion by human bronchial epithelial cells.

The airway epithelium, in addition to acting as a physicochemical barrier, plays a crucial role in initiating and augmenting pulmonary host defense mechanisms by synthesizing and releasing a variety of inflammatory mediators. The maintenance of leukocyte recruitment during inflammation requires intercellular communication between infiltrating cells, the endothelium, and the respiratory epithelium. Interleukin-8 (IL-8), 1 a member of the CXC family of chemokines, primarily functions as a potent chemoattractant and activator of neutrophils at sites of acute inflammation (1). Studies supporting the involvement of IL-8 in pulmonary inflammatory disorders have revealed elevated levels of IL-8 in the bronchoalveolar lavage fluid of patients with chronic obstructive lung disease, asthma, idiopathic pulmonary fibrosis, pulmonary sarcoidosis, and the acute respiratory distress syndrome (2)(3)(4)(5). Cultured airway epithelial cells also secrete significant amounts of IL-8 when exposed to cigarette smoke, diesel exhaust particles, ozone, hyperoxia, and viral infection as well as various endogenous mediators such as tumor necrosis factor-␣ and IL-1 (6 -11).
Considerable attention has been recently given to the diverse biological effects of sphingosine 1-phosphate (S1P), a bioactive sphingolipid derived from sphingomyelin that is released by activated platelets and multiple other cells in response to a wide array of stimuli (12,13). S1P activates intracellular signaling pathways by ligating, with high affinity, members of a G protein-coupled lysophospholipid receptor family. S1P specifically binds 5 of the 12 identified lysophospholipid receptors, S1P 1 , S1P 2 , S1P 3 , S1P 4 , and S1P 5 , also known as endothelial differentiation gene (EDG)-1, EDG-5, EDG-3, EDG-6 and EDG-8, respectively (14). Constitutively found in nanomolar to micromolar concentrations in human plasma and serum, S1P has been implicated in regulating cellular differentiation, chemotaxis, mitogenesis, and apoptosis (15)(16)(17)(18). Recently, S1P has been shown to be a complete angiogenic factor and the primary endothelial cell chemotactic factor in serum (19,20). In contrast to these vascular effects of S1P, reports in non-vascular tissue have described pro-inflammatory effects with IL-6 secretion in airway smooth muscle cells (21) and osteoblasts (22) and increased expression of IL-8 in ovarian cancer cells (23). Moreover, elevated S1P levels have been measured in the airways of asthmatic (but not control) subjects following segmental antigen challenge in association with inflammatory cell and protein influx (21). Thus, it appears that extracellular S1P is a mediator of inflammation by regulating proinflammatory cytokine and chemokine production.
Phospholipase D (PLD), an important effector enzyme in receptor-mediated signaling pathways, catalyzes the hydrolysis of the most abundant membrane phospholipid, phosphatidylcholine, and generates choline and phosphatidic acid (PA). Choline is rapidly phosphorylated to phosphorylcholine, which plays a role in cell proliferation (24). PA and its metabolites, lysophosphatidic acid and diacylglycerol, also act as second messengers in the regulation of secretion, mitogenesis, and proliferation (25). We and others (26 -28) have found that the S1P ligation of its G protein-coupled receptor results in a rapid activation of PLD in a variety of cell types. Furthermore, studies have suggested that PLD-derived PA is involved in thrombin-induced IL-6 synthesis in osteoblasts (29) and tumor necrosis factor-␣-induced NFB activation in a myeloblastic cell line (30), implicating PLD-derived PA as a signaling mediator acting upstream of transcription.
These studies support growing evidence that S1P is involved in the complex, tissue-specific process of inflammation. Because details and mechanisms of the inflammatory role of S1P remain unknown, we investigated the effect of S1P on bronchial epithelial PLD activation and its contribution to secretion of IL-8. Our results demonstrate that physiologically relevant concentrations of S1P markedly stimulated IL-8 secretion in a pertussis toxin (PTX)-sensitive manner, indicating the involvement of a G i -linked S1P receptor. Furthermore, we provide evidence that PLD, along with PKC and Rho, participates in the S1P-mediated signal transduction pathway resulting in airway epithelial cell secretion of IL-8.
Cell Culture-Beas-2B bronchial epithelial cells (passage 32; kindly provided by Dr. Sherer Sanders, Johns Hopkins University) were cultured in MEM supplemented with 5% FBS and penicillin/streptomycin/ amphotericin B. Cells were grown in 35-mm dishes at 37°C in 5% CO 2 , 95% air to 60 -70% confluence and serum-deprived by incubation for 12-18 h in MEM containing 0.1% FBS prior to IL-8 secretion experiments. For the PLD activity experiments, cells were effectively serumdeprived by prior overnight incubation in phosphate-free DMEM containing 2% FBS. All experiments with Beas-2B cells were conducted between 35 and 43 passages. Primary epithelial cells (HBEpC) were derived from surgical specimens of normal human bronchi from lung transplant donors and were isolated in the laboratory of Dr. Robert Schleimer (The Johns Hopkins University) by a modification of the method described by Schroth et al. (31) and Wu et al. (32). Briefly, airway specimens were rinsed with HEPES-buffered saline solution and then placed in dissociation solution consisting of F-12 media supplemented with penicillin (100 units/ml), streptomycin (100 g/ml), and Pronase (1 mg/ml; Roche Molecular Biochemicals) for 48 h at 4°C. After incubation, FBS (Sigma) was added to a final concentration of 20%, and epithelial cells were detached from the stroma by gentle agitation. The cells were collected by centrifugation, washed, and suspended in serumfree hormonally supplemented basal essential growth media (Bio-Whittaker, Walkersville, MD). All primary cell cultures were grown in basal essential growth media in 35-mm dishes at 37°C in 5% CO 2 , 95% air.
Immunocytochemistry-Beas-2B and HBEp cells grown on coverslips were rinsed with PBS and fixed in 3.7% formaldehyde for 15 min at room temperature followed by permeabilization with 0.5% Triton X-100 in PBS for 2 min. After washing 3 times with PBS, coverslips were incubated in blocking buffer (1% BSA in TBST) for 30 min. To detect specific S1P receptors, primary antibody incubations with either S1P 1 (1:200 dilution), S1P 2 (1:400 dilution), S1P 3 (1:1000 dilution), S1P 4 (1:100 dilution), or S1P 5 (1:100 dilution) were performed in blocking buffer for 1 h at room temperature. Cells were then washed 3 times with PBS and incubated with a 1:200 dilution of the appropriate secondary antibodies (anti-rabbit or anti-mouse) conjugated to Alexa Fluor 488 fluorescent dye in blocking buffer for 1 h at room temperature. After the final 3 washes with PBS, the coverslips were mounted using a commercial mounting medium (Kirkegaard & Perry Laboratories, Gaithersburg, MD). Analyses of immunofluorescent staining were performed using a Nikon TE200-S fluorescent microscope with ϫ60 objective lens and Hamamatsu digital camera and appropriate filter.
PLD Activation Measurement in Intact Cells-Beas-2B cells were incubated with [ 32 P]orthophosphate (5 Ci/ml) in phosphate-free Dulbecco's minimal Eagle's medium (DMEM) containing 2% FBS for 18 -24 h at 37°C in 5% CO 2 and 95% air incubator. The radioactive medium was removed by aspiration. Pretreatments were in MEM alone or with the specified agents for 1 h prior to stimulation, with the exception of PTX, which was 2 h. Cells were incubated in MEM containing 0.1% BSA with or without S1P or other agents at the indicated concentrations and in the presence of 0.05% butanol for the specified lengths of time. Incubations were terminated by the addition of 1 ml of methanol:HCl (100:1 v/v), and lipids were extracted in chloroform:methanol (2:1 v/v) (33). [ 32 P]Phosphatidylbutanol (PBt) formed as a result of PLD activation, and concomitant transphosphatidylation reaction (an index of in vivo PLD activation) was separated by TLC with the upper phase of ethyl acetate:2,2,4-trimethyl pentane:glacial acetic acid:water (65:10: 15:50 by volume) as the developing solvent system. Unlabeled PBt was added as a carrier during TLC separation of lipids and was visualized upon exposure to iodine vapors. PBt spots were marked, scraped, and radioactivity-associated PBt was determined by liquid scintillation counting. All values were normalized to 1 million disintegrations/min in total lipid extract and [ 32 P]PBt formed was expressed as dpm/dish or % control.
Measurement of IL-8 Secretion-Beas-2B cells were pretreated in MEM with or without PTX (40 ng/ml for 2 h), actinomycin D (1 g/ml for 1 h), BIM (1 M for 1 h), or Y27632 (indicated dose for 1 h) prior to stimulation. The pretreatment media were removed, and the cells were treated in MEM containing 0.1% BSA with or without S1P or other agonists at the indicated concentrations for the specified lengths of time. For the experiments involving 1-or 3-butanol (0.05%), pretreatments were for 15 min prior to stimulation, and the alcohol incubations were continued during stimulation with S1P. Cell supernatants were removed, centrifuged at 10,000 rpm for 5 min at 4°C, and frozen at Ϫ80°C for later analysis for IL-8 by ELISA, which was performed according to the manufacturer's instructions (BioSource International, Camarillo, CA).
Plasmid Construction and Transient Transfection-The human IL-8/luciferase (hIL-8/Luc) reporter plasmid contained the Ϫ162 to ϩ44 nucleotide region of the IL-8 promoter cloned into the poLUC reporter plasmid (generously provided by Dr. Allan R. Brasier, University of Texas Medical Branch, Galveston, TX, and Dr. Jeffrey Hasday, University of Maryland Medical System, Baltimore, MD) (34). Beas-2B cells at 50 -60% confluency were transfected in 35-mm diameter dishes by incubating overnight in 1 ml of Opti-MEM premixed with 3 l of FuGENE 6 and 1 g of hIL-8/Luc reporter plasmid. After 18 h, without changing the medium, the cells were treated with MEM containing 0.1% BSA with or without S1P (final concentration as indicated) for the specified lengths of time. After 6 h of treatment, the medium was removed; the cells were washed 3 times with PBS, and lysates were prepared to independently measure luciferase activity. Luciferase was then normalized to total protein concentrations.
Statistical Analysis of Data-S.D. for each data point was calculated from triplicate samples. Unless otherwise stated, data were subjected to one-way analysis of variance, and pairwise multiple comparison was done by Dunnett's method with p Ͻ 0.05.

Expression of S1P Receptors in Bronchial Epithelial
Cells-S1P binds multiple members of a G protein-coupled receptor family with tissue-specific receptor expression noted in various reports (12). S1P 1 , S1P 3 , and weak S1P 2 expression have been demonstrated in an immortalized nasal polyp epithelial cell line (35). S1P 4 , a recently cloned member of this lysophospholipid family of receptors that preferentially binds S1P, is expressed in lymphoid, hematopoietic tissue, and the lung (36). The other known S1P-specific receptor, S1P 5 , is widely expressed in the brain, as well as the lung, spleen, peripheral blood leukocytes, and aorta (37). We first examined which S1P receptors were present in Beas-2B and HBEp cells by Western blot analysis. As demonstrated in Fig. 1, S1P 1 (44 kDa), S1P 3 (45 kDa), S1P 4 (52 kDa), S1P 5 (42 kDa), and faint S1P 2 (45 kDa) receptor proteins were detected in both Beas-2B and HBEp cells. Complementing our Western blot results, immunocytochemical localization using S1P receptor antibodies on Beas-2B and HBEp cells is shown in Fig. 1. Immunofluorescent images shown are representative of the homogeneity of the monolayer visualized under the microscope. Overall, the distribution and localization of the individual S1P receptors were very similar for Beas-2B cells and HBEpC. In addition to localization at the plasma membrane, S1P 1 was detected in the cytoplasm and nuclear membrane and was generally excluded from the nucleus. S1P 2 was barely detectable as compared with the staining of the other S1P receptors. The distribution of S1P 3 appeared punctate and located in the plasma membrane and cytoplasm. Both S1P 4 and S1P 5 were primarily localized in the nucleus but were also observed in the cytoplasm and plasma membrane. Collectively, these findings indicate that the same S1P receptors are endogenously present in both Beas-2B and HBEp cells and that they are of similar quantities and distribution. S1P Activates PLD in Bronchial Epithelial Cells-The activation of PLD, a crucial signaling enzyme in secretory path-ways, by S1P has been demonstrated in endothelial cells (28) and other cell types, including the human lung adenocarcinoma A549 cell line (38). In the presence of 1-butanol (0.05%), PLD catalyzes a transphosphatidylation reaction to form PBt instead of PA. By using [ 32 P]PBt formation as an index of PLD activation, a 15-min stimulation of Beas-2B cells with S1P activated PLD in a dose-dependent fashion ( Fig. 2A). The activation was significant with 100 nM S1P (2.9-fold over control) increasing 4.5-fold with 1 M S1P, within physiological concentrations of S1P found in human serum (39). S1P (1 M) induced a rapid activation of PLD in Beas-2B cells (Fig. 2B), significantly increasing after 1 min of stimulation and reaching a maximum at 5 min (4.8-fold over control). This response was somewhat sustained for 30 min but declined after 1 h.
Because the S1P-specific receptors are capable of coupling to G proteins (40,41), we investigated the S1P-stimulated PLD activity after 2 h of pretreatment with PTX (40 ng/ml), which ADP-ribosylates and specifically inactivates G i proteins. PTX pretreatment nearly abolished the S1P-induced PLD activity (Fig. 3), suggesting the underlying signal transduction occurs via a G i -coupled S1P receptor. S1P Activates IL-8 in Bronchial Epithelial Cells-Extracellular S1P has been shown to activate IL-6 secretion in osteoblasts (22). More recently, it has been demonstrated that exogenous S1P can induce IL-6 secretion in human airway smooth muscle (21) and IL-8 expression in ovarian cancer cells (23). Increases in IL-8 production by airway epithelial cells have been linked to G protein-coupled receptor activation (42). We therefore examined the ability of S1P to modulate bronchial epithelial IL-8 secretion. Treatment of Beas-2B cells with S1P for 3 h resulted in a dose-dependent secretion of IL-8 ( Fig. 4A) with significant production at 100 nM (1.4-fold over control) and 4.6-fold over control at 1 M. As with the activation of PLD, IL-8 secretion induced by S1P occurred at physiological concentrations. The release of IL-8 into the medium by Beas-2B cells upon stimulation with S1P (1 M) significantly increased after 3 h (10-fold) and plateaued for 24 h (Fig. 4B). This time course of IL-8 secretion suggests that S1P-mediated signaling results in regulation of IL-8 at the transcriptional level. To test whether S1P stimulated IL-8 gene expression, the reporter gene luciferase under control of the Ϫ162 to ϩ44-nt region of the IL-8 promoter was transfected in Beas-2B cells. As shown in Fig. 5A, S1P treatment (1 M for 6 h) resulted in a 14-fold induction of IL-8 promoter-mediated luciferase activity. Pretreatment of Beas-2B cells for 1 h with actinomycin D (1 g/ml) (an antibiotic inhibitor of transcription) abolished the IL-8 response to S1P (Fig. 5B). These observations confirmed that the downstream effects of S1P occurred at the transcriptional level in inducing IL-8 secretion.
To determine whether the signal transduction involves coupling to G i proteins, we next studied the IL-8 response in the presence or absence of PTX (40 ng/ml). As shown in Fig. 6, PTX pretreatment for 2 h effectively blocked (95%) S1P-induced IL-8 secretion. These combined results strongly indicate that S1P-mediated pathways leading to both PLD activation and IL-8 secretion in Beas-2B cells are G i -dependent and that they are interrelated.
Role of PLD in S1P-induced IL-8 Secretion-The participation of PLD in protein trafficking and vesicle formation via ARF-dependent mechanisms has been well studied (43), but little is known about the association of PLD activation and transcriptional regulation. PLD-derived PA has been linked to AP-1 induction in T lymphocyte Jurkat cells (44) and NFB activation in human myeloblastic cells (30). Because the IL-8 promoter can be regulated by both AP-1 and NFB subunit binding (11), we hypothesized that PLD activation could lead to up-regulation of IL-8 transcription. The addition of 1-butanol (0.05%), which converts the PA formed to PBt by the PLDcatalyzed transphosphatidylation reaction, decreased S1P-induced IL-8 secretion by 60% in Beas-2B cells (Fig. 7), suggesting that PLD-generated PA mediates IL-8 expression. In the presence of the negative control, 3-butanol (0.05%), which cannot form PBt from PLD-generated PA (45), the S1P-mediated IL-8 response was almost identical to that of cells treated with S1P alone (Fig. 7).
To further support the finding that S1P-induced IL-8 secretion involves PLD-generated PA, we first examined the effect of treating the cells with TPA, a tumor promoter and potent activator of PKC that subsequently activates PLD (46). As shown in Fig. 8A, S1P (1 M) and TPA (25 nM) treatment for 5 min stimulated PLD activity ϳ9and 13-fold, respectively, over basal levels. Pretreatment of cells with the PKC inhibitor, bisindolylmaleimide (BIM) (1 M for 1 h), attenuated S1P-and TPA-mediated PLD activation by 35 and 75%, respectively (Fig. 8A). S1P (1 M for 3 h) and TPA treatment (25 nM for 3 h) resulted in 4.6-and 30-fold corresponding increases in IL-8 secretion, which were attenuated when pretreated with BIM (1 M for 1 h) by 45 and 90%, respectively (Fig. 8B). As shown in Fig. 9, TPA-mediated IL-8 secretion was modestly but significantly reduced by 29% in the presence of 1-butanol (0.05%), whereas 3-butanol (0.05%) had no significant effect. These data provide strong evidence that PKC and PLD-generated PA participate in the S1P signaling cascade leading to bronchial epithelial cell production of IL-8.
Involvement of Rho in S1P-induced IL-8 Secretion-Receptor-mediated activation of PLD has been shown to be via a PKC-dependent process in many systems (43). Other studies have identified roles for the monomeric GTP-binding protein, Rho, in PLD activation (47,48) and IL-8 secretion (49). To address the involvement of Rho in S1P-mediated PLD activation and IL-8 secretion in Beas-2B cells, we used the Rho kinase inhibitor, Y27632. In Fig. 10A, S1P-induced PLD activation was moderately reduced (24%) in cells pretreated with Y27632 (10 M for 1 h). Furthermore, pretreatment with the Rho kinase inhibitor (1 h) dose-dependently attenuated S1Pinduced IL-8 secretion, 24% at 1 M, 50% at 10 M, and 90% at 20 M (Fig. 10B). These results indicate that Rho has a role in

S1P-induced IL-8 secretion in Beas-2B cells and suggest that
PLD is a component of this arm of the S1P-mediated signaling. DISCUSSION The balance between the protective, physiological inflammatory response and injurious, pathological inflammation involves the coordination of multiple extra-and intracellular mediators. The airway epithelium, a critical first line environmental defense, plays a pivotal role in this response through its reaction to and secretion of these messengers in order to provide the necessary intercellular regulatory signals. One such paracrine messenger, IL-8, a well established potent neutrophil chemoattractant and activator, is elevated in the airways of patients with a variety of respiratory disorders and exposures. Another extracellular first messenger, S1P, promotes growth factor-like responses in cells by ligating S1P-specific receptors (50). Recent evidence indicates that S1P can serve as an inflammatory effector molecule by inducing IL-6 secretion in airway smooth muscle cells (21) and osteoblasts (22) and IL-8 expression in ovarian cancer cell lines (23). Although many current studies have focused on the proliferative, survival, morphogenic, and angiogenic properties of S1P (12, 51), inves- tigations of inflammatory functions with respect to the respiratory epithelium are lacking. Our findings presented here demonstrate that in bronchial epithelial cells constitutively expressing S1P receptors, extracellular S1P stimulates the production of IL-8. S1P may have intracellular sites of action, but they have not been thoroughly identified. In agreement with a growing body of literature, our data suggest that S1P-induced IL-8 secretion follows the binding of a G protein-coupled receptor. The expression of S1P receptors changes during development and differentiation; thus tissue distribution occurs with heterogeneity (16,53). Recent reports have described the in vivo physiological significance of S1P receptors by using mice with disrupted S1P receptor genes. S1P 1 -null mice are embryonically lethal due to a deficiency of vascular smooth muscle cells leading to incomplete vascular maturation (54), indicating the requirement of S1P 1 for mouse development. Although S1P 3 -null mice are viable, fertile, and have normal development without phenotypic abnormalities, examination of null mice fibroblast signal transduction pathways revealed significant loss of S1P-induced phospholipase C activation (55). Similarly, S1P 2 -null mice display no gross anatomical defects, yet have spontaneous seizures from 3 to 7 weeks of age accompanied by hyperexcitablity of neocortical neurons (56). Future studies with these mutant mice should provide valuable information about the role of S1P receptors in various disease processes. S1P 1 , S1P 3 , S1P 4 , S1P 5 , and weak S1P 2 have been detected in human lung tissues (37,57), similar to our detected pattern of expression in Beas-2B and HBEp cells (Fig. 1). Interestingly, the immunocytochemical analysis revealed that the S1P receptors apparently have differential cellular localization depending on the receptor type. We find that S1P receptors can be detected not only on the plasma membrane but also distributed throughout the cytoplasm, intranuclear, or nuclear membrane. S1P 1 phosphorylation and internalization have been demonstrated with S1P ligation of the receptor (58,59). Our studies were conducted with the cells in an unstimulated state. The predominant nuclear distribution of S1P 4 and S1P 5 suggests that these receptors may be intracellular targets of the S1P, generated via sphingosine kinase, following agonist stimulation (60). With respect to the signaling pathway presented here, the possibility of the involvement of more than one S1P receptor in S1P-induced IL-8 secretion cannot be excluded. A detailed determination of native S1P receptor localization could aid in identifying unique intracellular sites of action.
Lysophospolipid receptors exhibit differential coupling to heterotrimeric G proteins (61), making functional analysis complex. For instance, S1P 1 and S1P 4 couple directly to the G i pathway, whereas S1P 2 and S1P 3 stimulate the G i , G q , and G 12/13 pathways (50). S1P 5 activates both G i and G 12/13 proteins (62). Complexities are encountered when attempting to differentiate S1P receptor/G protein signal transduction, including the lack of specific pharmacologic inhibitors and truly negative heterologous expression systems with most cell lines expressing at least one species of S1P receptor or exhibiting intrinsic biological responses to S1P. Nonetheless, few studies have investigated S1P receptor-mediated signaling and function in non-transformed cells that endogenously express S1P receptors. Our results with PTX pretreatment indicate that S1Pinduced PLD activation and IL-8 production are dependent on G i protein-mediated signaling responses (Figs. 3 and 6).
In addition to its role in vesicular trafficking (63), PLD is involved in cellular proliferation and mitogenesis (25). Our laboratory has reported the activation of PLD by S1P in endothelial cells (28), whereas others have shown that S1P induces PLD activity in CFNPE9o Ϫ and A549 airway epithelial cell lines (35,38). Physiologic concentrations of S1P (100 nM to 1 M) rapidly activated PLD in Beas-2B cells, with maximal activity at 5 min after stimulation (Fig. 2, A and B). The rapid rate of S1P-induced PLD activity implies a transduction of signals with minimal intermediates. PLD generates the important second messenger, PA, which in turn regulates multiple intracellular signaling responses. Stimulation of PLD has been implicated in NFB activation in fibroblasts (30) and activator protein-1 (AP-1) activation in Jurkat T cells (44). NFB and AP-1 are the principal transcription factors that regulate IL-8 gene expression (11).
The expression and secretion of IL-8 can be induced by diverse inflammatory stimuli and require de novo synthesis (11). In cells transfected with the hIL-8/Luc reporter plasmid, S1P treatment resulted in a strong induction of luciferase activity (Fig. 5A), indicating that S1P stimulated IL-8 gene transcription. The Ϫ162to ϩ44-nt region of the IL-8 gene promoter contains binding sites for AP-1 (between Ϫ127 and Ϫ119 nt) and NFB (between Ϫ80 and Ϫ70 nt) (51), suggesting that extracellular S1P stimulates one or both of these transcription factors to regulate IL-8 gene transcription. Consistent with this known pathway of regulation, inhibition of transcription by pretreatment of Beas-2B cells with actinomycin D, thereby blocking transcription, abolished the S1P-induced IL-8 production (Fig. 5B).
Besides catalyzing the hydrolysis of phosphatidyl choline to generate PA, in the presence of a primary alcohol, PLD carries out a transphosphatidylation reaction forming a phosphatidyl alcohol from PA. This unique enzymatic feature of PLD is FIG. 11. Proposed signaling mechanisms involved in S1P-induced IL-8 secretion. S1P ligates its receptor(s) on the epithelial cell surface, thereby activating G protein-coupled signaling cascades. The resulting IL-8 production is primarily dependent on G i protein, PKCand PLD-mediated pathways (heavy arrows), but is also controlled in part by Rho (light arrows). PLD activation appears common and downstream in both pathways. PCh, phosphatidylcholine. utilized to differentiate the activity of PLD from that of PLC and to inhibit specifically downstream signaling events governed by PLD-generated PA. Our data showing a significant (60%) decrease in S1P-induced IL-8 secretion in the presence of low concentrations of 1-butanol (Fig. 7) suggest that the PA formed by PLD is an important signaling regulator in this pathway. This finding was further strengthened by using a negative control and showing a lack of inhibition in cells pretreated with a tertiary alcohol, 3-butanol (Fig. 7). This is also supported by our observation that TPA-stimulated IL-8 secretion, via PKC-dependent PLD activation, was also attenuated by 1-butanol (Fig. 9).
Given that a myriad of signals are simultaneously activated by G protein coupling, it is not surprising that blocking PLD activity did not result in total attenuation of S1P-mediated IL-8 secretion. Many agonists whose receptors are linked to heterotrimeric G proteins regulate PLD via activation of PKC and/or small G proteins of the Rho family (48,64,65). Our results indicate that Rho is involved in S1P-induced PLD activation and IL-8 secretion (Fig. 10). The lower attenuation by Rho kinase inhibition, as compared with PKC inhibition (Fig. 8), of both PLD activation and IL-8 secretion suggests that the PKCmediated pathways dominate. S1P receptor-coupling to G 12/13 , which S1P 1 and S1P 4 are not known to couple (61,66), may be the major route of Rho activation (67). Contrary studies have shown S1P 1 receptor signaling to be Rho-dependent in cells overexpressing S1P 1 (40) and suggested the involvement of Rho in G i protein-mediated pathways (51). Similar to these studies, our results show that PKC, and to a lesser extent Rho, are regulators of S1P-mediated IL-8 secretion, with PLD activation playing a central role in the signal transduction. The near total inhibition of S1P-induced IL-8 secretion by PTX (Fig. 6) indirectly supports the notion that G i proteins can activate Rho. Additionally, our findings are in agreement with previous reports (49,68) demonstrating and supporting the participation of PKC and Rho in agonist-mediated NFB-dependent gene expression and IL-8 transcription.
In endothelial cells, S1P is capable of activating p38 mitogenactivated protein kinase (MAPK) (19), and p38 MAPK regulates PLD (69). Furthermore, we have shown in Beas-2B cells that S1P-induced p38 MAPK phosphorylation is 1-butanolinsensitive, whereas ERK1/2 phosphorylation is attenuated with 1-butanol. 2 These two MAPKs, particularly p38, have been shown to be involved in NFB activation (69) and IL-8 synthesis (49,52,70). It stands to reason that parallel and concomitant signaling events occur when S1P binds to its receptor(s). Some of these numerous signals likely converge on one key component such as NFB-mediated IL-8 transcription, whereas others take a divergent pathway that affects a seemingly unrelated target. Future studies aimed at the elucidation of specific S1P 3 S1P-receptor 3 G protein pathways and downstream signaling events leading to the production of IL-8, as well as other cytokines, are necessary to gain a better understanding of lysophospholipid regulation of inflammation.
In summary, our findings demonstrate that PLD is involved in S1P-induced IL-8 secretion in airway epithelial cells, a pathway and function not described previously (Fig. 11). The IL-8 production elicited by S1P appears to occur via a G i proteincoupled pathway and to also involve PKC and Rho. Although a paucity of studies exist that describe the participation of S1P in airway inflammation, our data indicate that S1P may augment local inflammatory reactions. The development and utilization of S1P inhibitors, S1P receptor antagonists, and animal models of lung injury, as well as the determination of the sources of airway S1P, will enable clarification of the role of S1P under in vivo and in vitro inflammatory conditions.