Lysophosphatidic Acid Activates NF-κB in Fibroblasts

Lysophosphatidic acid (LPA) is a growth factor that exerts a number of well characterized biological actions on fibroblasts and other cells. In the present study, we investigated the possibility that LPA activates the transcription factor NF-κB. NF-κB is a target of cytokines, but its activation by other classes of agonists has raised considerable interest in the control of processes such as inflammation and wound healing through varied mechanisms. We find that LPA causes a marked activation of NF-κB in Swiss 3T3 fibroblasts as determined by the degradation of IκB-α in the cytosol and the emergence of κB binding activity in nuclear extracts. The EC50 for activation of NF-κB is 1–5 μm, a range similar to that reported for reinitiation of DNA synthesis and activation of the serum response element. Activation of NF-κB is attenuated by pertussis toxin and inhibitors of protein kinase C, and it is completely blocked by the Ca2+ chelator BAPTA-AM. The combination of phorbol ester and thapsigargin promotes an activation comparable with that of LPA. Activation by LPA is additionally inhibited by tyrphostin A25 but not genistein or AG1478, indicating a selective utilization of protein-tyrosine kinases, and by certain antioxidants, implying a role for reactive oxygen species. The activation is also inhibited by tricyclodecan-9-yl-xanthogenate (D609), implying a requirement for hydrolysis of phosphatidylcholine. The data demonstrate the utilization of multiple pathways in the activation of NF-κB by LPA, not inconsistent with the relevance of several families of GTP-binding regulatory proteins.

Lysophosphatidic acid (LPA) is a growth factor that exerts a number of well characterized biological actions on fibroblasts and other cells. In the present study, we investigated the possibility that LPA activates the transcription factor NF-B. NF-B is a target of cytokines, but its activation by other classes of agonists has raised considerable interest in the control of processes such as inflammation and wound healing through varied mechanisms. We find that LPA causes a marked activation of NF-B in Swiss 3T3 fibroblasts as determined by the degradation of IB-␣ in the cytosol and the emergence of B binding activity in nuclear extracts. The EC 50 for activation of NF-B is 1-5 M, a range similar to that reported for reinitiation of DNA synthesis and activation of the serum response element. Activation of NF-B is attenuated by pertussis toxin and inhibitors of protein kinase C, and it is completely blocked by the Ca 2؉ chelator BAPTA-AM. The combination of phorbol ester and thapsigargin promotes an activation comparable with that of LPA. Activation by LPA is additionally inhibited by tyrphostin A25 but not genistein or AG1478, indicating a selective utilization of protein-tyrosine kinases, and by certain antioxidants, implying a role for reactive oxygen species. The activation is also inhibited by tricyclodecan-9-yl-xanthogenate (D609), implying a requirement for hydrolysis of phosphatidylcholine. The data demonstrate the utilization of multiple pathways in the activation of NF-B by LPA, not inconsistent with the relevance of several families of GTP-binding regulatory proteins.
Lysophosphatidic acid (LPA 1 ; 1-acyl-2-lyso-sn-glycero-3phosphate) is a naturally occurring, water-soluble glycerophospholipid that exhibits striking hormone-and growth factor-like activities (1,2). Synthesized and released by platelets, LPA represents a major bioactive constituent of serum, and its actions on fibroblasts, endothelial cells, and smooth muscle cells in particular suggest roles in wound healing among other events. Indeed, LPA acts on a large number of cells to achieve a broad range of immediate and long lasting effects. Specific responses to LPA include changes in cell shape and tension, chemotaxis, proliferation, and differentiation.
The molecular actions of LPA have been characterized best in rodent fibroblasts, where at low concentrations (i.e. 10 -100 nM) LPA stimulates phosphoinositide hydrolysis (3) and promotes the Rho-dependent formation of stress fibers and focal adhesions (4). The stimulation of phosphoinositide hydrolysis is thought to occur through the GTP-binding regulatory protein (G protein) G q . The formation of stress fibers and focal adhesions is consistent with activation of Rho through G 12 or G 13 (5). One or a combination of these G proteins is also responsible for the protein tyrosine phosphorylation elicited by LPA (6 -8). LPA additionally inhibits adenylyl cyclase, an action achieved through the pertussis toxin (PTX)-sensitive G protein G i (9). LPA uses G i , moreover, to activate Ras, Raf, and the extracellular signal-regulated kinases (ERKs) ERK1 and ERK2 (10,11). The activation of ERKs (11), and presumably the inhibition of adenylyl cyclase (3), occurs at concentrations of LPA comparable with those stimulating phosphoinositide hydrolysis and cytoskeletal changes. At higher concentrations (i.e. 5-70 M), LPA promotes reinitiation of DNA synthesis in quiescent fibroblasts (9). Whether G proteins are sufficient for this action is unclear, but the sensitivity of the phenomenon to PTX implies that G i represents at least one necessary input. The need for high concentrations of LPA in this context may relate to a requirement for more persistent signaling and/or engagement of other receptors or pathways. Micromolar concentrations of LPA also promote arachidonic acid formation, a second phase of inositol phosphate accumulation (PTX-insensitive), and activation of serum response factor (9,12,13). Receptors that recognize LPA are poorly characterized; however, several have been identified that conform to the seven-transmembrane domain motif characteristic of G protein-coupled receptors (GPCRs) (14 -17).
NF-B (nuclear factor-B) is the prototype of a family of dimers whose constituents are members of the Rel family of transcription factors (18). In most types of cells, NF-B is present as a heterodimer comprising p50 (NF-B1) and p65 (RelA). NF-B is normally retained in the cytosol in an inactive form through interaction with IB inhibitory proteins. Release of NF-B for translocation to the nucleus and interaction with cognate DNA sequences is accomplished through a signal-induced phosphorylation and subsequent degradation of IB. Originally described as a necessary element for expression of the immunoglobulin gene in mature B cells, NF-B is now recognized to be an important transcriptional regulatory protein in a variety of cell types (19).
The binding of agonists to certain GPCRs promotes activation of NF-B. Agonists include serotonin (working through the 5-HT 1A receptor) (20), platelet-activating factor (21), thrombin * This work was supported by National Institutes of Health Grant GM51196. 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.
‡ To whom correspondence should be addressed: Dept. of Pharmacology, University of Pennsylvania School of Medicine, 3620 Hamilton Walk, Philadelphia, PA 19104-6084. Tel.: 215-898-1775; Fax: 215-573-2236; E-mail: manning@pharm.med.upenn.edu. 1 The abbreviations and trivial names used are: LPA, lysophosphatidic acid; ERK, extracellular signal-regulated kinase; G protein, GTPbinding regulatory protein; GPCR, G protein-coupled receptor; PKC, protein kinase C; PMA, phorbol-12-myristate-13-acetate; PTX, pertussis toxin; TNF␣, tumor necrosis factor-␣; EGF, epidermal growth factor; D609, tricyclodecan-9-yl-xanthogenate; BAPTA-AM, 1,2-bis(o-aminophenoxy)ethane-N,N,NЈ,NЈ-tetraacetic acid tetra(acetoxymethyl) ester. (22), and bradykinin (23). That GPCRs are linked to NF-B is particularly significant, since these receptors are widely distributed, the actions of NF-B are notable, and the coincident activation of NF-B and other GPCR-regulated transcription factors can provide unique forms of transcriptional regulation. Because LPA exerts a wide range of actions in part or entirely through GPCRs, and because NF-B is especially relevant to inflammation and wound healing, we instituted efforts here to understand whether LPA promotes the activation of NF-B. We explored the possible relationship between LPA and NF-B in fibroblasts and the mechanisms by which this relationship is established.
Cell Culture-Swiss 3T3 mouse embryo fibroblasts (a gift from Dr. E. Rozengurt, Imperial Cancer Research Fund, London, UK) were maintained at 37°C under a humidified atmosphere of 10% CO 2 in Dulbecco's modified Eagle's medium containing 10% (v/v) fetal calf serum, supplemented with penicillin (100 units/ml) and streptomycin (100 g/ml). For all experiments, 1 ϫ 10 6 cells were subcultured into 10-cm tissue culture plates (Nunc). After 48 h, the medium was replaced with 6 ml of Dulbecco's modified Eagle's medium containing 1% fetal calf serum and antibiotics, and the cells were incubated for an additional 18 h. LPA (prepared as a stock of 1 mg/ml in phosphate-buffered saline containing 10 mg/ml essentially fatty acid-free bovine serum albumin (Sigma)) and/or other reagents or vehicles were added to achieve the concentrations specified.
Nuclear Extract Preparation-Nuclear extracts were prepared by the method of Dignam et al. (24) with minor modifications. Following incubation with LPA or other reagents, cells were washed twice in ice-cold phosphate-buffered saline, harvested, and resuspended in 400 l of hypotonic buffer (10 mM HEPES (pH 7.9), 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride). After 10 min on ice, 30 l of Nonidet P-40 (10% (v/v)) was added with mixing for 2 s. The nuclei were pelleted by centrifugation at 20,000 ϫ g for 10 s. The supernatant was removed, and the nuclei were resuspended in hypertonic buffer (20 mM HEPES (pH 7.9), 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 0.1 mM phenylmethylsulfonyl fluoride) and shaken for 45 min at 4°C. The samples were centrifuged at 20,000 ϫ g for 30 s, and the supernatants (nuclear extracts) were saved. Protein concentrations were determined using the Bradford method (Bio-Rad).
Electrophoretic Mobility Shift Assay-Electrophoretic mobility shift assays were performed using a double-stranded oligonucleotide containing a consensus B binding site (5Ј-AGTTGAGGGGACTTTC-CCAGGC-3Ј; the underlined sequence represents the consensus B region), which was end-labeled with [␥-32 P]ATP and T4 polynucleotide kinase. Nuclear extracts (2.5 g of protein) were incubated in 10 mM Tris-HCl (pH 7.9), 50 mM NaCl, 1 mM EDTA, 10% glycerol, 0.15 mg/ml poly(dI-dC), and 20 -30 fmol of 32 P-labeled oligonucleotide (50,000 -100,000 cpm) in a total volume of 15 l at room temperature for 10 min. The reaction mixture was then subjected to electrophoresis in a 5% polyacrylamide slab gel containing 50 mM Tris, 380 mM glycine, and 2 mM EDTA, pH 8. The gels were dried under vacuum for analysis by autoradiography (overnight exposure) or PhosphorImager analysis. For competition studies, nuclear extracts were incubated prior to the addition of labeled oligonucleotide for 10 min at room temperature with unlabeled oligonucleotide, unlabeled oligonucleotide containing a G 3 C substitution in the B binding motif (5Ј-AGTTGAGGCGACTTTC-CCAGGC-3Ј), or an unlabeled oligonucleotide containing the consensus binding site for Sp1 (5Ј-ATTCGATCGGGGCGGGGCGAGC-3Ј). For supershift analysis, nuclear extracts were incubated with approximately 2 g of antibodies specific for p65 (RelA) or nonimmune goat IgG for 30 min at 4°C in the presence of radiolabeled oligonucleotide prior to electrophoresis. The results shown in all figures are representative of at least three experiments.
Western Blot Analysis-Following exposure to LPA with or without cycloheximide as specified, cells were washed with ice-cold phosphatebuffered saline and lysed in 10 mM Tris-HCl (pH 7.6), 5 mM EDTA, 50 mM NaCl, 30 mM sodium pyrophosphate, 50 mM NaF, 100 mM Na 3 VO 4 , 0.5% Triton X-100, 10 mg/ml leupeptin, 10 mg/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride. Lysates were clarified by centrifugation at 20,000 ϫ g for 15 min at 4°C. Supernatants were collected and subjected to SDS-polyacrylamide gel electrophoresis (11% acrylamide). Protein was transferred to nitrocellulose membrane and probed with polyclonal rabbit anti-human IB-␣ (0.4 mg/ml) detected subsequently by chemiluminescence using a donkey anti-rabbit IgG conjugated with horseradish peroxidase and luminol as recommended by the manufacturer (ECL Western; Amersham Pharmacia Biotech).

RESULTS
The possibility that LPA activates the transcription factor NF-B was investigated in Swiss 3T3 fibroblasts, which have been used extensively in studies of agonists, including LPA, linked to changes in cell morphology and reinitiation of DNA synthesis (4,25,26). We examined first the extent to which LPA promotes degradation of IB-␣, an inhibitory protein whose proteolysis would precede the translocation of NF-B to the nucleus. As shown in Fig. 1 (left panel), LPA caused a transient degradation of this protein. Levels of IB-␣ decreased slowly following introduction of LPA, reaching a minimum at 40 -60 min, and increased thereafter to near control values. The time-dependent resynthesis of IB-␣ is a common finding in cytokine action (27) and appears to reflect activation of the IB-␣ gene by NF-B as part of a feedback loop (28). To circumvent resynthesis of IB-␣, we also evaluated degradation of this protein in the presence of cycloheximide. As expected, degradation of IB-␣ under this condition, where protein synthesis is blocked, was complete.
A more direct evaluation of NF-B activation was carried out by electrophoretic mobility shift assays. The data in Fig. 2 demonstrate that LPA promotes a time-and concentration-dependent appearance of a factor(s) within nuclear extracts that binds to an oligonucleotide probe containing an NF-B binding site. The proinflammatory cytokine TNF␣ also promotes the appearance of this factor. The relevant protein-DNA complex was evident as a band of radioactivity (denoted by an arrow) positioned above two less prominent bands. This band, but not the other two, was supershifted with a p65 (RelA)-directed antibody (Fig. 3, top panel), confirming the identity of the induced factor as NF-B. The nature of the protein-DNA interaction was evaluated further in competition experiments (Fig.  3, bottom panel), where the 32 P-labeled oligonucleotide was found to be displaced by unlabeled oligonucleotide. The same unlabeled oligonucleotide, but containing a mutation in the B site, and an altogether unrelated oligonucleotide (containing an Sp1 binding site) did not displace the 32 P-labeled oligonucleotide. The EC 50 for LPA based on the intensity of the shifted band was 1-5 M, and the time required for full development of the response was 40 -60 min (Fig. 2). The response was transient, as the level of shifted oligonucleotide began to decrease by 3 h (not shown).
The extent to which the G protein G i might contribute to the activation of NF-B was assessed with PTX, which suppresses activation of G i by GPCRs. Pretreatment of cells with PTX attenuated LPA-induced activation of NF-B by approximately 60% (Fig. 4, top panel). Efforts to enhance the attenuation by manipulating pretreatment time and concentration of PTX were unsuccessful. That the actions of LPA typically assigned to G i are not completely suppressed by PTX, for example ERK activation and reinitiation of DNA synthesis, is not without precedent (9 -11).
A downstream target for both G i and G q is the phosphoinositide-specific phospholipase C-␤, whose activation results in recruitment of PKC and mobilization of Ca 2ϩ . Overnight treatment of cells with a high concentration of PMA (1 M) to induce down-regulation of classical and novel forms of PKC suppressed activation of NF-B by 70 -80% (not shown), as did Ro-31-8220 (Fig. 4, middle panel), which inhibits all forms of PKC. Other inhibitors of PKC, including bisindolylmaleimide I (not shown) and calphostin C, inhibited activation of NF-B nearly as well. Activation of NF-B was completely suppressed by pretreatment of cells with the cell-permeable Ca 2ϩ chelator BAPTA-AM (Fig. 4, bottom panel).
Given the apparent requirements for PKC and intracellular Ca 2ϩ , we tested whether the activation of PKC and/or mobilization of Ca 2ϩ might be sufficient to activate NF-B. Some degree of activation was achieved with PMA, but not the extent observed with LPA (Fig. 5). Only a small degree of activation was achieved with thapsigargin, moreover, an inhibitor of the endoplasmic reticular Ca 2ϩ -ATPase that causes a timedependent increase in cytosolic Ca 2ϩ . However, the combination of thapsigargin and PMA activated NF-B ultimately to an extent somewhat greater than that achieved with LPA.
Because considerable attention has been devoted to the utilization of the EGF receptor and other tyrosine kinases by GPCRs, including the one or more receptors that mediate the actions of LPA (29, 30), we tested inhibitors of different tyrosine kinases for their effect on LPA-induced activation of NF-B. Genistein, which inhibits a wide range of tyrosine kinases, was without effect (Fig. 6). Tyrphostin AG1478, a relatively specific inhibitor of the EGF receptor tyrosine kinase, caused some degree of inhibition but only at very high concentrations (16 M is shown; concentrations normally used to inhibit the EGF receptor are 0. 125-1 M (29, 31)). At concentrations less than 5 M, AG1478 had no effect. In contrast, tyrphostin A25, like genistein regarded as a general inhibitor of tyrosine ki-   Chen et al. (33) demonstrated that LPA stimulates reactive oxygen species production in HeLa cells and that antioxidants inhibit LPA-stimulated MAP kinase kinase activity. We therefore evaluated the effects of antioxidants on the activation of NF-B. As shown in Fig. 7, N-acetylcysteine completely inhibited LPA-induced activation of NF-B. Pyrrolidinedithiocarbamate was similarly effective. Dimethyl sulfoxide achieved a less extensive, but still notable, inhibition. Ascorbic acid and dithiothreitol were without effect. The activation of NF-B was also highly sensitive to diphenyleneiodonium, an inhibitor of flavanoid-containing enzymes such as NADPH oxidase.
The tricyclodecan xanthogenate D609 inhibits the hydrolysis of phosphatidylcholine at the level of a phosphatidylcholinespecific phospholipase C-like enzyme and/or phospholipase D (34 -36), among perhaps other actions (37), and has been used to explore signaling pathways utilized by growth factors, TNF, and GTPase-deficient G protein subunits. Schü tze et al. (35), for example, found that D609 inhibits the activation of NF-B by TNF␣, and Wadsworth et al. (38) demonstrated that this compound inhibits Na ϩ /H ϩ exchange stimulated by ␣ 12 . We found here that D609 inhibits quite effectively the activation of NF-B by LPA (Fig. 8). The EC 50 was about 3 g/ml, and the maximum degree of inhibition was greater than 90%. Potassium ethylxanthate had no effect. D609 was not nearly as potent an inhibitor of TNF␣'s activation of NF-B as it was of LPA's. Inhibition in the case of TNF␣ occurred only at concentrations of D609 exceeding 50 g/ml. The activation of NF-B by LPA was thus selectively inhibited by D609. DISCUSSION That proinflammatory cytokines such as interleukin-1 and TNF␣ activate NF-B has long been appreciated, and the sequence of events by which the activation occurs is now emerging (39,40). Of perhaps equal significance, however, is the fact that agonists working through GPCRs can also activate NF-B (20 -23). In the work described here, we have focused on LPA. LPA has a rich and important biology; it induces a number of cells to proliferate and others to differentiate and, at a molecular level, works through one or more G proteins to stimulate MAP kinases, phospholipid metabolism, and cytoskeletal rearrangement. Activation of NF-B clearly constitutes an additional action of LPA, and one to be reconciled in any setting of transcription relevant to this agonist.
LPA triggers a pronounced activation of NF-B, as ascertained by the degradation of IB-␣ and the emergence of B binding activity in nuclear extracts. Based on supershift experiments, the activated form of NF-B contains the p65 (RelA) subunit. The other component may be p50 (NF-B1), given the widespread occurrence of NF-B as a p50-p65 heterodimer, but this remains to be determined. Dimeric complexes of NF-B that contain p65 function as strong activators of gene expression (41). The transience of NF-B activation suggests that IB-␣ alone of the inhibitory proteins is the target of LPA's action (28).
The concentration of LPA supporting activation of NF-B (EC 50 ϭ 1-5 M) is higher than that reported to inhibit adenylyl cyclase (3), activate ERKs (10), elevate IP 3 and Ca 2ϩ (3), and stimulate formation of stress fibers and focal adhesions (4) but is similar to that needed for activation of serum response factor (12), initiation of a "second" phase of inositol phosphate production (9), and reinitiation of DNA synthesis (9). It is tempting to speculate that activation of NF-B by LPA, like that of serum response factor, may function as a counterpart to proliferative signaling. Mayo et al. (42) have reported that activation of NF-B by oncogenic Ras is required for progress toward cell transformation and, in particular, that NF-B prevents a Ras-induced apoptosis that would otherwise abrogate transformation. It is not implausible that mitogens such as LPA, as they drive replication, similarly use NF-B to foil any tendency of the cell to enter into an apoptotic program. Alternatively, the activation of NF-B may play a more direct role in proliferative signaling, for example through activation of genes immediately relevant to DNA synthesis.
Because the roles of G proteins in the phenomena induced by LPA are not completely established, a basic question is whether LPA utilizes G proteins to achieve its activation of NF-B. That PTX attenuates activation of NF-B would suggest a role for G i . The participation of this G protein would not be surprising; NF-B can be activated by oncogenic Ras (43), and both Ras and ERKs are activated by LPA through a pathway(s) at least partly sensitive to PTX (10,11,44). pp90 rsk1 lies downstream of ERKs and has been implicated in the phosphorylation of IB-␣ (45,46). Yet, the activation of G i would seem to occur at concentrations of LPA lower than those required for activation of NF-B (3, 10, 11). Thus, while G i may contribute to the activation of NF-B, it may not provide a sufficient stimulus. The actions of PTX, moreover, must be qualified, since PTX conceivably affects the activation of NF-B by LPA indirectly, for example through elevations in cAMP or through actions not related specifically to ADP-ribosylation. G q may well provide an additional and/or independent input. As implied above, the stimulation of inositol phosphate produc- tion by LPA occurs in two phases (3,9). Some production of inositol phosphates is observed at low concentrations of agonist, but the preponderance is achieved at micromolar concentrations. The latter (at least) is insensitive to PTX (9). It is conceivable that low concentrations of LPA activate G i and that higher concentrations activate G q . G q may be responsible for the activation of protein kinase C and mobilization of Ca 2ϩ of sufficient magnitude and/or duration to bring, together with signals from G i , activation of NF-B to fruition. That the combination of PMA and thapsigargin achieves activation of NF-B supports the notion that G q might even be sufficient. However, PMA and thapsigargin would exert more potent and long lasting actions than those achieved by agonist-activated G q . With respect to the involvement of PKC, we have used both downregulation and several pharmacological inhibitors to implicate a role for this enzyme(s). That Ro-31-8220 can activate JNK is a potentially confounding issue, but bisindolylmaleimide I lacks this attribute (47).
The possibility that Rho is engaged by G 12 or G 13 in the activation of NF-B by LPA is most intriguing. LPA appears to activate Rho (4), and Rho when overexpressed as a wild-type or constitutively active protein activates NF-B (48). The activation of Rho by LPA no doubt proceeds through G 12 and/or G 13 . The GTPase-deficient forms of ␣ 12 and ␣ 13 , for example, cause a Rho-dependent formation of stress fibers and focal adhesions, as does LPA (5). Labeling of G proteins with [ 32 P]GTP azidoanilide, moreover, reveals activation of G 12 and G 13 , together with G i and G q , by LPA (31). Tyrphostin A25 has been demonstrated previously to prevent the induction of stress fiber and focal adhesion assembly by the GTPase-deficient form of ␣ 13 (31), as it does that by LPA at a step upstream from Rho (32). Of interest, we find here that A25 prevents LPA-induced activation of NF-B. Thus, the activation of NF-B by LPA may well involve an A25-sensitive activation of Rho. Rho has recently been demonstrated to be required in the activation of NF-B by bradykinin in human epithelial cells (49). How a Rho-dependent pathway integrates with those pathways sensitive to inhibitors of PKC and intracellular Ca 2ϩ fluxes, D609, and PTX remains to be determined. In a departure from the results presented for cytoskeletal changes (31), we do not have firm evidence for the utilization of the EGF receptor by LPA. AG1478, an inhibitor of EGF receptor autophosphorylation, had only a modest impact at high concentrations on the activation of NF-B. Genistein, which would block phosphorylation of the receptor by Src among other kinases, had no effect.
The relevance of reactive oxygen intermediates to the activation of MAP kinases and NF-B has generated considerable interest (33, 50 -52). N-Acetylcysteine has been used extensively in this regard and was found here to effectively inhibit the activation of NF-B by LPA, as did pyrrolidinedithiocarbamate and dimethyl sulfoxide. The nature and source of the reactive oxygen species is unclear, although the inhibition achieved with diphenyleneiodonium might suggest the involvement of NADPH oxidase. Of interest, several reports find reactive oxygen species to be involved in activating components of the ERK activation cascade, ostensibly through phosphorylation of the EGF receptor (50,53). For reasons outlined above, we suspect that the EGF receptor is not a participant in the activation of NF-B, and we know that ERK is activated at concentrations of LPA well below those required for activation of NF-B. It is nevertheless intriguing that the reactive oxygen species-dependent phosphorylation of the EGF receptor reported in HeLa cells is implied to occur at only higher (i.e. micromolar) concentrations of LPA (53). Perhaps, then, the formation of reactive oxygen species will emerge as the most closely correlated requirement for NF-B activation, SRF activation, and reinitiation of DNA synthesis.
The activation of NF-B by LPA is exceptionally sensitive to inhibition by the tricyclodecan xanthogenate D609. Sensitivity to this compound has been used in other contexts to suggest the relevance of a phosphatidylcholine-specific phospholipase C (34,35). More recent reports suggest that phospholipase D, too, can be inhibited by D609 (36,54), perhaps at higher concentrations. Our use of D609 was prompted by the observation that D609 inhibits the activation of NF-B by TNF␣ (35), as it does that of sphingomyelinase and of Raf by oncogenic Ras (34). We found that D609 indeed inhibited the activation of NF-B by TNF␣ but only at high concentrations (Ͼ50 g/ml), consistent with the observation of Schü tze et al. (35). In contrast, the amount of D609 required to inhibit activation of NF-B by LPA was considerably lower; almost complete suppression could be achieved at concentrations less than 10 g/ml. D609 at these lower concentrations has been argued to inhibit a phosphati- Swiss 3T3 fibroblasts were preincubated as indicated for 30 min with D609 or the inactive analog potassium ethylxanthate (PEX) and then exposed to 40 M LPA (or vehicle) or 30 ng/ml TNF␣ for 40 min. Nuclear extracts were prepared for electrophoretic mobility shift assays. dylcholine-specific phospholipase C-like activity selectively (36). The most obvious contribution of a phosphatidylcholinespecific phospholipase C would be the activation of Raf (34) and hence ERKs and pp90 rsk by LPA-activated Ras. However, low concentrations of D609, perhaps through inhibition of a phosphatidylcholine-specific phospholipase C but conceivably other means, can cause hyperphosphorylation and concomitant release of Raf from membrane as a form of negative modulation (37). We therefore cannot rule out the possibility that D609 has actions beyond phosphatidylcholine hydrolysis. It is of considerable interest that ␣ 12 , which can activate Rho, utilizes a D609-sensitive step in the regulation of Na ϩ /H ϩ exchange (38). The possibility therefore exists that ␣ 12 utilizes the same step in the activation of Rho, Raf, and/or PKCs (55), leading to activation of NF-B.
Whether any or all of the receptors for LPA so far cloned mediate the activation of NF-B represents a subject of interest. Receptors include PSP24, Edg2 (Vzg-1), and Edg4 (14 -17). Edg2 and Edg4, at least, are linked to the activation of the serum response element (16,17), and the activation through Edg4 was demonstrated to be partly sensitive to PTX and to the inhibition of Rho. Edg2 and Edg4 therefore represent reasonable candidates for linking LPA to NF-B. Regardless, the activation of NF-B, together with the means by which the activation is coordinated with other signals elicited by LPA (e.g. activation of ERK and serum response factor), must now be considered in any response to LPA of significant duration and/or explicitly involving gene expression.