INTRODUCTION

Lysophosphatidylcholine (lyso-PC)¹ is a natural phospholipid that can be generated intracellularly by the action of phospholipase A₂ on membrane phosphatidylcholine, the most abundant cellular phospholipid (1). Strong evidence supports a role for lyso-PC in the processes of atherogenesis, inflammation, and wound healing (1-7). The lyso-PC content of atherosclerotic arteries is severalfold higher than that of normal vessels (2, 3). It constitutes up to 40% of total lipid in atherogenic lipoproteins such as oxidatively modified low density lipoprotein and -migrating very low density lipoprotein (8). Lyso-PC has been identified as an essential component responsible for some biological activities of these lipoproteins in vivo, e.g. the chemotactic effect on human monocytes (6) and mitogenic action on macrophages (9). Additionally, lyso-PC is locally generated by the action of secretory phospholipase A₂ in wounds and inflammatory lesions (1), suggesting its involvement in response to injury and other stresses.

The addition of lyso-PC to cultured cells can transcriptionally up-regulate the expression of a variety of genes including cell adhesion molecules (intercellular adhesion molecule 1 and vascular cell adhesion molecule 1) (10), growth factors (platelet-derived growth factors A and B and heparin-binding epidermal growth factor) (11, 12), and vasoprotective enzymes such as nitric-oxide synthase (13, 14) and cyclooxygenase-2 (15). In vascular smooth muscle, lyso-PC has been shown to induce vascular relaxation (16, 17) and to stimulate cell proliferation (18, 19). It has also been reported that lyso-PC significantly potentiated protein kinase C (PKC)-mediated cellular responses such as primary T-lymphocyte activation (20, 21) and HL-60 cell differentiation into macrophages (22). Since actions of many extracellular agonists are associated with the activation of membrane phospholipase A₂ and the subsequent accumulation of lyso-PC (23-29), it is conceivable that lyso-PC may act as a second messenger, transducing signals elicited from membrane receptors. This is consistent with experiments showing that lyso-PC and phospholipase A₂ under certain circumstances have similar effects when incubated with cultured cells (21, 22).

Despite the multiple biological activities of lyso-PC, no underlying signal transduction mechanisms have been revealed. In the present study, we report that exogenous lyso-PC induces AP-1 DNA binding and transcriptional activity and activation of the c-Jun NH₂-terminal kinase (JNK1). Although lyso-PC may activate other signal transduction pathways, the stimulated JNK and AP-1 activity defined by this study probably plays a role in mediating the response to lyso-PC.

EXPERIMENTAL PROCEDURES

Palmitoyl-lyso-PC, palmitoyl-lyso-PA, and oleoyl-lyso-PA were purchased from Avanti Polar Lipids. Solutions of these phospholipids were prepared fresh in phosphate-buffered saline containing 1.5% fatty acid-free bovine serum albumin. The phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) was obtained from Sigma. cDNA of the human c-jun gene was a gift of Dr. J. Woodgett (University of Toronto, Toronto, Ontario, Canada) and cDNA of the human glyceraldehyde-3-phosphate dehydrogenase gene was from ATCC. [³²P]ATP, [³²P]dCTP, and goat anti-rabbit or anti-mouse antibodies coupled to horseradish peroxidase were from Amersham Corp. Antibodies against JNK1, ERK2, or the c-Jun and AP-1 consensus sequence were obtained from Santa Cruz Laboratories. Poly(dI·dC)·poly(dI·dC) was from Pharmacia Biotech Inc. Cell culture reagents including medium and calf serum or fetal bovine serum were from Life Technologies, Inc. Unless otherwise stated, other chemicals used in this study were from Sigma.

Swiss 3T3 and HeLa cells were obtained from ATCC and maintained as recommended by the supplier. Rat-1 fibroblasts were kindly provided by Dr. W. H. Moolenaar (The Netherlands Cancer Institute, Amsterdam, The Netherlands) and routinely grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Bovine aortic endothelial cells (BAEC) were isolated and phenotypically characterized as reported (30). Homogeneous populations of cells were grown in dishes coated with 0.2% bovine gelatin and serially propagated in RPMI 1640 medium with 15% calf serum. Cells between passages 3 and 8 were used for the experiment. Prior to treatment with various stimuli, Swiss 3T3 cells, Rat-1 cells, HeLa cells, and BAEC were made quiescent by growing to confluence without changing the medium for at least 5 days (31) or by serum starvation.

Nuclear extracts were prepared as described by Sadowski and Gilman (32). Briefly, following treatment with lyso-PC, TPA, or vehicle, cells were rinsed once with ice-cold phosphate-buffered saline, once with phosphate-buffered saline containing 1 mM Na₃VO₄ and 20 mM NaF, and once with a hypotonic buffer (33). The cells were then lysed with hypotonic buffer plus 0.2% Nonidet P-40, and lysates were briefly centrifuged. The pelleted nuclei were resuspended in hypotonic buffer with 420 mM NaCl and 20% glycerol added and rocked gently for 30 min at 4 °C, followed by a 20-min spin (14,000 rpm) in a microcentrifuge.

Gel shift assay of AP-1 binding was performed as described by Zohn et al. (33) with the exception that a shorter AP-1 probe (21 base pairs) was used. The specificity of binding to the ³²P-labeled AP-1 probe was confirmed by experiments showing inhibition of the binding activity by the cold (unlabeled) oligonucleotide. Reaction mixes were run on 5% native polyacrylamide gels and autoradiographed.

A plasmid (pBL3 × TRECAT2) containing the bacterial chloramphenicol acetyltransferase gene under the control of the minimal herpes simplex virus thymidine kinase gene promoter linked to three consensus AP-1 binding sites (3 × TRE) was provided by Dr. P. Chiao (University of Texas M. D. Anderson Cancer Center, Houston, TX). The chloramphenicol acetyltransferase gene was replaced by the luciferase gene isolated from pGL2-Basic (Promega) to generate pBL3 × TRELuc. The luciferase reporter vectors containing the truncated human collagenase (Col) gene promoter sequence (73/+63Col-Luc, 60/+63Col-Luc) (34) or c-jun gene promoter sequence (79Jun-Luc) (34, 35) were gifts of Dr. M. Karin (University of California, San Diego) and B. Su (University of Texas M. D. Anderson Cancer Center). In 79Jun-Luc-AP1mu, the TRE site in 79Jun-Luc was mutated (5-GTGACATCAT-3 to 5-GATCCACCAT-3) using polymerase chain reaction. Swiss 3T3 cells were transfected with lipofectamine (2 µg of plasmid/60-mm dish) according to the protocol of the manufacturer (Life Technologies, Inc.). The transfected cells were refed with fresh medium the next day and grown to confluence and near quiescence (4 days) before treatment with lyso-PC, lyso-PA, or TPA in serum-free medium. Cells were incubated with lyso-PC or lyso-PA for 1.5 h and then switched to plain medium until 6 h had passed. This is designed to minimize possible metabolism of the phospholipids. TPA was incubated with cells for 6 h before harvest. Cell extracts were prepared and assayed for luciferase activity using a kit from Promega. The luciferase activity was normalized for protein concentrations of extracts. For each plasmid in each experiment, transfection efficiency among randomly chosen dishes was found to be very consistent, with an S.D. value less than 10% of means.

Cells were treated with lyso-PC, TPA, lyso-PA, or vehicle for the indicated times, washed with cold phosphate-buffered saline, and lysed at 4 °C in a buffer containing 25 mM HEPES (pH 7.4), 0.3 M NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM dithiothreitol, 1.0% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 20 mM -glycerophosphate, 1 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride, 20 µg/ml aprotinin, and 20 µg/ml leupeptin. JNK was immunoprecipitated from cleared lysates by incubation with a rabbit anti-JNK antibody for 2 h at 4 °C followed by a 1-h incubation with protein A-Sepharose beads (Pharmacia). Immunocomplexes were then washed and analyzed for kinase activity as described by Coso et al. (36) using glutathione S-transferase-c-Jun as substrate (37) (a gift of Dr. J. Woodgett). MAPK (ERK1 and ERK2) activity was determined by an in-gel assay with myelin basic protein as substrate as described previously (38).

Lysates containing equal amounts of cellular protein or immunoprecipitates were analyzed by Western blotting after SDS-polyacrylamide gel electrophoresis, transfer to Immobilon (polyvinylidene difluoride), and incubation with primary antibodies. Immunocomplexes were visualized with an enhanced chemiluminescence detection kit (Amersham) using horseradish peroxidase-conjugated secondary antibodies.

Total cellular RNA was extracted from cultured cells using the guanidinium isothiocyanate-phenol chloroform method. Poly(A) RNA was isolated directly from cells by using the FastTrack 2.0 kit from Invitrogen. RNA samples (15 µg of total cellular RNA or 5 µg of mRNA) was size-fractionated by formaldehyde/agarose gel electrophoresis, stained with ethidium bromide, and transferred to N⁺ hybrid nylon. RNA was immobilized by UV cross-linking and then prehybridized and hybridized to ³²P-labeled cDNA probes in 50% formamide, 6 × SSC, 10 × Denhardt's solution, 10 mM EDTA, 0.1% SDS, and 150 µg/ml denatured salmon sperm DNA. Quality and comparable loading of RNA were confirmed by rehybridization of nylon membranes to the ³²P-labeled cDNA of glyceraldehyde-3-phosphate dehydrogenase or -actin.

RESULTS

AP-1 transcription factor is an important regulator of gene expression in response to growth factors, cytokines, carcinogens, and many other stimuli (39). To assess the potential effect of lyso-PC on AP-1 DNA binding activity, Swiss 3T3 fibroblasts were incubated with palmitoyl lyso-PC, and the AP-1 activity of nuclear extracts was examined by a gel shift assay as described under "Experimental Procedures." Lyso-PC treatment increased AP-1 DNA binding activity in a concentration- and time-dependent manner (Fig. 1, A and B). The increased AP-1 activity was detectable at 10 µM lyso-PC. At 20 µM, AP-1 activity reached a level only slightly lower than that induced by 1 h of TPA treatment. At 40 µM of lyso-PC, peak stimulated AP-1 activity was increased, as was the duration of activation. Some cytotoxicity of lyso-PC was, however, observed at this concentration. After incubation with 20 µM of lyso-PC, increased AP-1 DNA-binding activity was detected after 30 min, peaked at 1.5 h, remained elevated for 2 h, and declined thereafter (Fig. 1B). Compared with lyso-PC, TPA, a potent AP-1 activator, induced stronger and more prolonged activation of AP-1 DNA binding (Fig. 1C).

The effects of lyso-PC on AP-1-mediated transcriptional activity in Swiss 3T3 cells were examined using luciferase constructs containing the truncated human collagenase promoter (73Col-Luc) and the c-jun gene promoter (79Jun-Luc) (Fig. 1D). Lyso-PC treatment of cells induced a moderate stimulation of luciferase activity with 73Col-Luc (4-5-fold). This effect was not seen in cells transfected with 60Col-Luc, a construct with the TRE deleted from the promoter (34). With the 79Jun-Luc reporter, up to 20-fold stimulation was observed in lyso-PC-treated cells. This activity was essentially eliminated when the vector was mutated at the AP-1 site (79Jun-Luc-AP1mu), confirming that the stimulation of transcription is specific and AP-1-dependent. Lyso-PA, a bioactive phospholipid that can be potentially generated from breakdown of lyso-PC, did not significantly induce transcription from 79Jun-Luc.

In these experiments, TPA has been included as positive control. However, it had only a weak effect (2-3-fold), compared with that of lyso-PC, on transcriptional activity of the 73Col-Luc and 79Jun-Luc reporters. The finding differs from previous measurements of AP-1 activity in other cell types that relied on synthetic promoters containing multiple AP-1 sites. However, using another construct with thymidine kinase promoter linked to three TREs (3 × TRE-TK-Luc) under the same experimental conditions, we found that TPA was a fairly potent activator of AP-1 transcriptional activity (10-fold) (Fig. 1D). Consistent with our observations here, the differential effects of TPA on multiple TRE elements and the physiological promoters have been also noticed by others (34). It seems that, depending on cell types and experimental conditions, increased AP-1 complex binding to a single TRE element is not sufficient for optimal activation of transcription. Cooperation of the other cis-elements, which are not present in the 73Col-Luc or 79Jun-Luc, may be required for the full response of the endogenous genes to TPA. Furthermore, the difference between lyso-PC and TPA indicates that they stimulate AP-1 transcriptional activity through independent mechanisms.

Activation of PKC can potently stimulate cellular AP-1 activity resulting from dephosphorylation of c-Jun serine residues near the DNA binding domain as well as increased c-Fos synthesis (39, 40). Some previous studies have suggested that actions of lyso-PC depend on intact PKC activity (41, 42). As the results in Fig. 1D have suggested that lyso-PC-induced AP-1 transcriptional activity may be independent of PKC, we then examined whether PKC is required for lyso-PC-stimulated DNA-binding activity. Swiss 3T3 cells were pretreated for 1.5 h with a potent and specific PKC inhibitor, GF 109203X (5 µM), which interacts with the ATP-binding site of PKC (43). In these pretreated cells, a complete inhibition of PKC activity was achieved, as reflected by the failure of TPA to stimulate any increase in AP-1 DNA binding activity. In contrast, lyso-PC was capable of inducing a full magnitude of increase in the activity (Fig. 2). This clearly established that lyso-PC-induced AP-1 activity is independent of PKC. Furthermore, as demonstrated in Fig. 3, lyso-PC did not activate MAPK (ERK1 and ERK2), providing evidence that lyso-PC does not induce PKC activation, which would otherwise cause activation of MAPK (33) (see Fig. 3).

Growth factors and many other extracellular agonists induce AP-1 activity through activation of MAP kinases (ERK1 and ERK2) and the subsequent augmentation in c-Fos expression (44). We examined the possible contribution of ERK1 and ERK2 to the increased AP-1 activity in lyso-PC-treated cells. The phosphorylation state of ERK2 was assessed in a gel mobility shift assay by Western blot analysis of cell lysates with anti-ERK2 antibody. It has been well documented that phosphorylation of ERK2 leads to a shift in mobility, which is correlated with an increased kinase activity (44, 45). As shown in Fig. 3, no mobility shift of ERK2 was observed in cells treated with different concentrations of lyso-PC (1, 10, and 20 µM) for different intervals (5, 10, 15, 20, and 25 min). In contrast, such a mobility shift was readily detected in the cells treated with known MAPK stimulators, TPA and lyso-PA. Since lyso-PC-induced AP-1 activity was maximal between 1 and 2 h, examination of ERK2 phosphorylation was extended to 0.5, 1, 1.5, and 2 h with no change in the mobility seen (data not shown). Similarly, in-gel kinase assays (38) did not detect any increased ERK1 or ERK2 activity in lyso-PC-treated cells in contrast to TPA- or lyso-PA-stimulated cells, which showed marked increase in ERK1 and ERK2 activity (data not shown).

Distinct members of the MAPK family, the JNKs or stress-activated protein kinases (SAPKs), have been recently characterized (37, 46). JNK can be activated by a variety of stimuli, including inhibitors of protein synthesis such as cycloheximide and anisomycin, inflammatory cytokines such as tumor necrosis factor- and interleukin-1, UV irradiation, heat shock, and other cellular stresses (37, 46). It has been shown that JNK activation is responsible for phosphorylating the transactivating domain of the c-Jun protein in vivo (46), and in turn, phosphorylated c-Jun homodimers have potent AP-1 activity, which regulates the expression of a number of genes including c-jun itself (35, 39). Therefore, we tested the effect of lyso-PC on JNK1 activation by an in vitro kinase assay using the c-Jun N terminus (residues 1-79) fused to glutathione S-transferase as a substrate (37). In contrast to ERK, JNK activity was greatly stimulated by lyso-PC in a dosage- and time-dependent manner (Fig. 4). Lyso-PA at the same range of concentrations did not induce JNK activation (data not shown). JNK activity was detectable at 10 µM of lyso-PC, and enhanced activity was seen with 20 and 40 µM. At a fixed concentration of lyso-PC (20 µM), JNK activity was detected beginning at 30 min, reached a peak at 1 h, decreased from 1.5-2 h, and disappeared by 4 h. The kinetics of JNK activation preceded that of AP-1 DNA binding activity, compatible with JNK activation being causally related to the increase in AP-1 activity.

As further evidence of JNK1 activation, gel mobility shift analyses of JNK1 protein showed the presence of a phosphorylated form of JNK1 in lyso-PC-treated cells (Fig. 4). The ratio of the phosphorylated JNK1 relative to the unphosphorylated form was consistent with the observed increases in kinase activity.

We also determined whether lyso-PC-induced JNK activation required PKC activity. Swiss 3T3 cells were chronically treated with TPA (1 µM) for 24 h, which depleted functional PKC activity, as reflected by abolishment of PKC-dependent activation of ERK-2 by further TPA treatment (data not shown). Lyso-PC stimulated JNK1 phosphorylation and the corresponding kinase activity in TPA-pretreated cells as efficiently as in the cells without TPA pretreatment (Fig. 4C). We conclude that PKC is not required for stimulation of JNK activation by lyso-PC unless some TPA-insensitive isotypes such as PKC- are involved (47).

To study the effect of lyso-PC on other types of cells in addition to Swiss 3T3, we examined JNK1 status in Rat-1 fibroblasts, BAEC, and human HeLa epithelial cells following treatment with lyso-PC. In each of these cell types, JNK1 was strongly activated by lyso-PC as reflected by high magnitude of phosphorylation (Fig. 5). In contrast, as with Swiss 3T3 cells, lyso-PC treatment failed to induce ERK2 phosphorylation in Rat-1 fibroblasts, BAEC, and HeLa cells.

The c-jun gene contains two AP-1 binding sites in its promoter (35, 39). The ability of lyso-PC to activate the truncated c-jun gene promoter (Fig. 1D) and to stimulate JNK (Fig. 4) predicts that lyso-PC would activate transcription of the endogenous c-jun gene. Northern blot analysis of total cellular RNA was conducted to test this prediction. Lyso-PC indeed induced a transient increase in c-jun mRNA expression at concentrations that stimulated JNK and AP-1 activity (Fig. 6). Since newly synthesized c-Jun protein would potentially contribute to the increase in AP-1 DNA binding activity, it is reasonable that lyso-PC-induced AP-1 DNA binding activity (Fig. 1B) peaked approximately 30 min later than did JNK activation (Fig. 4B) and c-jun mRNA expression (Fig. 6B).

Since JNK activation has been linked to not only c-Jun transactivation but also the production of AP-1 proteins (c-Jun and c-Fos) (34, 48-50), it is conceivable that the observed increases in DNA binding as well as transcriptional activity in lyso-PC-treated cells are subsequent to JNK activation. However, it is also possible that the AP-1 activity may be derived mainly from a JNK-independent stimulation of c-Jun and c-Fos protein production. If this is the case, lyso-PC would lose its stimulatory role on an AP-1-responsive promoter when protein synthesis is inhibited. We assessed the possibility by measuring the effect of lyso-PC on AP-1-mediated transcription in the presence of protein synthesis inhibitors. We chose emetine and puromycin instead of cycloheximide, which itself is a strong activator of JNK (37). Swiss 3T3 cells were transfected with 79Jun-Luc. Emetine was added to culture 1.5 h before treatment with lyso-PC for 1 h. Lysates and mRNA were prepared from parallel cultures. As shown in Fig. 7, emetine itself did not activate JNK. Nor did it influence lyso-PC-induced JNK activation. Protein synthesis was efficiently blocked by emetine, as reflected by the loss of c-Jun protein accumulation in lyso-PC-treated cells in comparison with the corresponding cells in the absence of emetine (Fig. 7). However, such inhibition of protein synthesis did not prevent lyso-PC from inducing transcription from 79Jun-Luc. The -fold increase in luciferase mRNA level following 1 h of stimulation with lyso-PC is only slightly lower than that in control cells without protein synthesis inhibition. This minor difference could be simply due to the higher background in emetine-treated cells. Alternatively, AP-1 protein synthesis may, to a certain degree, potentiate lyso-PC-induced transcription from the AP-1-responsive promoter. Thus AP-1 protein synthesis is not essential for the initial AP-1 transcriptional activity in lyso-PC-treated cells. Similar results have been obtained using another protein synthesis inhibitor, puromycin (data not shown). We have also made stably transfected clones carrying 73Col-Luc or 79Jun-Luc. When protein synthesis was inhibited in these clones, stimulation of the luciferase mRNA following a 1-h treatment of lyso-PC was largely maintained (data not shown). These observations indicate that lyso-PC stimulates AP-1-mediated transcription mainly through activation of preexisting AP-1 protein(s), although a hypothesized JNK-independent pathway leading to the production of AP-1 proteins in lyso-PC-treated cells cannot be ruled out and may contribute to the observed increase in AP-1 DNA binding. It does not seem that such a contribution to the DNA binding activity is either essential or sufficient for the functional activation (transcription). This is also in agreement with our results with TPA, which induces AP-1 DNA binding but did not optimally activate transcription from 73Col-Luc or 79Jun-Luc (Fig. 1).

DISCUSSION

Because it is implicated in wound healing, inflammation, and atherogenesis (1-7), lyso-PC is believed to act as a regulator of functions/dysfunction of fibroblasts and endothelial cells in vivo and in vitro. However, little is known as to how lyso-PC elicits its actions on cells. In this report, we show that lyso-PC is a potent stimulator of JNK1 and AP-1 activity. Because JNK activation has been linked to c-Jun transactivation and AP-1 protein synthesis, our findings are compatible with a model in which lyso-PC stimulates AP-1 activity through activation of JNK, although other JNK-independent mechanisms causing increased production of AP-1 proteins may exist. In addition to c-Jun, TCF, ATF2, and p53 have recently been identified as substrates of JNK1 (49-51). Thus the JNK-AP-1 cascade defined by this study could contribute to lyso-PC-induced expression of a number of different genes that contain regulatory sites for these transcriptional factors in their promoters. In this context, it is not surprising that the AP-1 transcription factor has been implicated in the regulation of expression of intercellular adhesion molecule 1, which is also stimulated by lyso-PC (10, 52).

The rapidly growing literature on JNK indicates that it is most potently activated by pathways other than those utilized by traditional growth factor receptor tyrosine kinases. The reported activators including UV irradiation, tumor necrosis factor-, and interleukin-1, as well as stresses such as hydrogen peroxide and heat shock, involve signal transduction pathways that fail to fully activate ERK (37, 53). The pathway to JNK activation by these diverse stimuli seems to be more complicated than the Ras-Raf-KEK/MKK-ERK pathway, which is common among growth factor signaling (54). A parallel module, Rac/Cdc42-PAK-MEKK1-SEK/MKK4-JNK/SAPK, has been postulated for JNK activation from extracellular stimuli (36, 54, 55). There exist, however, Rac, Cdc42-independent, and even SEK-independent pathway(s) leading to activation of JNK from diverse stimuli, suggesting that JNK can be activated through multiple pathways (36, 55, 56).

In our study, we have not defined the upstream signals leading to JNK activation by lyso-PC. Previous studies with human monocytes and a lymphoblastic cell line have implicated the generation of diacylglycerol via a metabolic pathway involving (lyso)phospholipase C, in the action of exogenous lyso-PC (57, 58). As an intracellular second messenger, diacylglycerol may conceivably stimulate PKC activation (59). Intact PKC activity has been shown to be required for some actions of lyso-PC, e.g. stimulation of superoxide anion production in vascular tissue (41) and modification of G-protein signaling in endothelial cells (42). Despite these observations, our demonstration of PKC-independent stimulation of JNK and AP-1 activity indicates that PKC is not required for the signaling process leading to JNK and AP-1 activation by lyso-PC. In addition, the failure of lyso-PC to stimulate ERK1 and ERK2, downstream targets of PKC, in Swiss 3T3 cells suggests that lyso-PC does not activate PKC. This seems to contradict observations described above that suggested the importance of PKC in some effects of lyso-PC (41, 42, 57, 58). These previous studies, however, only showed the dependence upon the intact PKC activity for specific actions of lyso-PC without directly demonstrating activation of PKC by lyso-PC. It could be that basal PKC activity is sufficient under these circumstances or, alternatively, that lyso-PC-mediated activation of PKC is cell type-specific.

It has been reported that lyso-PC treatment of certain types of cells can stimulate a biphasic increase in [Ca²⁺]_i, resulting from transient mobilization from intracellular Ca²⁺ stores and sustained influx from an extracellular source (60, 61). A potential role for this change in [Ca²⁺]_i to JNK activation by lyso-PC is compatible with a recent report describing calcium-dependent, angiotensin II-induced JNK activation in the GN4 rat liver epithelial cell line (33). As suggested by the authors, however, this could be a cell type-specific pattern. Furthermore, increases in [Ca²⁺]_i are not sufficient to induce JNK activation, since calcium ionophore (A23187) failed to stimulate JNK activity in lymphocytes as well as several fibroblastic and epithelial cell lines (34).

JNK is potently activated by stress such as UV irradiation, hyperosmosis, heat shock, and protein synthesis inhibition (37, 46, 62, 63). The mechanism by which stress culminates in JNK activation remains largely unknown. It appears that these diverse stresses stimulate JNK activation through distinct cellular sensors. For instance, UV-induced, but not heat shock-mediated JNK activation is dependent on membrane-associated components and free oxygen radicals (64). Although lyso-PC at concentrations (10 and 20 µM) that sufficiently activated JNK did not cause evident cytotoxicity, the possibility cannot be excluded that lyso-PC, as a polar lipid, may cause other alterations on the membrane bilayer. This may create a stress-like condition that ultimately leads to JNK activation.

Since other similar phospholipids, such as lyso-PA and platelet-activating factor have specific G-protein-coupled receptors on the cell membrane (65, 66), lyso-PC may also interact with a specific membrane receptor. It has been suggested that lyso-PC may act as an agonist for the thromboxane and platelet-activating factor receptors (17, 67). We are currently testing whether lyso-PC activates JNK through binding to these receptors. Another possibility is that lyso-PC may enter cells and serve as an intracellular second messenger with JNK as a downstream target. This hypothesis has been strengthened by a series of studies demonstrating that lyso-PC was generated intracellularly through activation of phospholipase A₂ following stimulation of tyrosine kinase receptors by macrophage colony-stimulating factor (23), and stimulation of G-protein-coupled receptors by thrombin (24, 25), bombesin (26), interferon- (27), or angiotensin (28, 29). The identification of potential intracellular effectors of lyso-PC should help in elucidating the mechanism(s) for the multiple biological functions of this phospholipid.