Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Extracellular signal-regulated kinases (ERK) are a group of mitogen-activated protein kinase (MAPK) that phosphorylates proteins at a motif of Ser/Thr-Pro (Widmann et al., 1999). Of the ERK isoforms identified in mammalian cells, ERK1 and ERK2 have been studied most extensively. A number of proteins have been shown to be phosphorylated by ERK1/2, such as p90^rsk S6 kinase, phospholipase A₂, and transcription factors (Widmann et al., 1999). Most of the studies to date indicated that ERK1/2 are involved in proliferation, differentiation, and apoptosis (Widmann et al., 1999). For instance, in megakaryocytes, activation of ERK1/2 by thrombopoietin is associated with endomitosis and an increase of megakaryocyte-specific antigens (Fichelson et al., 1999; Rojnuckarin et al., 1999), but recent experiments using kinase inhibitors indicated that ERK1/2 may also participate in proliferation induced by stem cell factor and interleukins (Fichelson et al., 1999).

In contrast to the cytokine receptors, ERK1/2 activation by G protein-coupled receptors in megakaryocytes is rarely documented, and the effects of these receptors on megakaryocytopoiesis remain to be determined. In K562 leukemia cells, treatment of protein kinase C activators resulted in an increase of ERK1/2 activity and expression of megakaryocyte markers (Racke et al., 1997; Whalen et al., 1997). These findings imply that G protein-coupled receptor agonists may contribute to the production and commitment of megakaryocyte lineage. We are interested in the signaling processes of G protein-coupled receptors in hematopoietic cells and have used human erythroleukemia (HEL) cells to characterize the effects of E-series prostaglandins and thrombin on phosphoinositide hydrolysis, phospholipase D, adenylyl cyclase, and calcium mobilization (Brass et al., 1991; Wu et al., 1991, 1992). The occurrence and interaction of various receptors and G proteins have been investigated in this megakaryocytic cell line as well (Motulsky and Michel, 1988; Michel et al., 1989; Schwaner et al., 1992; Feoktistov et al., 1994; Baltensperger and Porzig, 1997; Keffel et al., 1999), making it an attractive cellular model for G protein-coupled receptor signaling in megakaryocytes. To better understand the signaling processes of G protein-coupled receptors in megakaryocyte-like cells, the goals of the present study were dual. Our first goal was to determine whether ERK1/2 would be stimulated by receptors coupled to various types of G proteins in a native system in which receptors, G proteins, and signaling pathways have been characterized. In addition, we were interested in exploring the regulatory mechanisms and cellular functions of ERK1/2 activation by G protein-coupled receptors in cells of hematopoietic origin. In this study, we have extended previous observations by examining the types of receptors to which ERK1/2 are linked in HEL cells and further exploring the effects of these receptors on proliferation. Our results demonstrate that ERK1/2 activation is selective for receptors coupled to specific types of G proteins, involves multiple pathways, and is dissociated from mitogenesis in HEL cells.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Anti-ERK2 antibody was obtained from Upstate Biotechnology (Lake Placid, NY), and anti-ERK1/2 antibody was obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Anti-phospho-ERK1/2 antibody was purchased from New England Biolabs (Beverly, MA). Alkaline phosphatase- and horseradish peroxidase-conjugated secondary antibodies were purchased from Bio-Rad (Hercules, CA). Fetal bovine serum and RPMI 1640 medium were purchased from Invitrogen (Carlsbad, CA). Bisindolylmaleimide I was obtained from Calbiochem (La Jolla, CA). BAPTA/AM was obtained from Molecular Probes (Eugene, OR). U0126 was obtained from Promega (Madison, WI). Reagents for gel electrophoresis and Western blotting were purchased from AMRESCO (Solon, OH). Nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate were obtained from Roche Molecular Biochemicals (Indianapolis, IN). Chemiluminescence reagents were purchased from Amersham Biosciences (Piscataway, NJ). [³H]Thymidine was from PerkinElmer Life Sciences (Boston, MA). All other agents were purchased from Sigma (St. Louis, MO).

Cell Culture. HEL cells were from the American Type Culture Collection (Manassas, VA) and grown at 37°C in RPMI 1640 medium supplemented with 10% (v/v) fetal bovine serum and 1 mM glutamine. The cell density was maintained between 2 × 10⁵ and 1 × 10⁶ cells/ml in suspension culture by dilution with fresh medium.

Assay of ERK1/2 Activation. Stimulation of ERK1/2 was determined by tyrosine phosphorylation in most of the experiments, using polyclonal antibodies specific for the phosphorylated form of ERK1/2 (New England Biolabs). A typical assay is as follows. HEL cells (5 × 10⁵ cells/ml) were incubated in RPMI 1640 in the absence of serum overnight and treated at 37°C with vehicle or various agents for the appropriate times. For those experiments requiring pretreatment, cells were incubated with vehicle or inhibitors for the times indicated before stimulation. After stimulation, cells were rapidly pelleted and lysed by sample buffer. Total cell lysates (100-150 µg of protein) were separated by gel electrophoresis (10% acrylamide gel) and transferred onto nitrocellulose membrane. The membrane was blotted with anti-phospho-ERK1/2 antibodies (1:1000) overnight and then with conjugated secondary antibodies, followed by visualization with colorimetric agents (bromochloroindolyl phosphate and nitro blue tetrazolium) or with chemiluminescence.

View larger version (42K): [in this window] [in a new window]   Fig. 1.   Expression and phosphorylation of ERK1/2 in HEL cells. A, total cell lysates were separated by electrophoresis and ERK1/2 expression was analyzed by Western blotting. A polyclonal antibody was used to detect both the ERK1 and ERK2 (lane 1). In lane 2, the blot was probed with a monoclonal antibody specific for the ERK2. Numbers shown on the right indicate the molecular mass (kDa) of protein standards. B, serum-starved cells were stimulated with or without 1 µM PGE2 for the times indicated. Phosphorylation of ERK1/2 was determined by Western blotting using a polyclonal antibody specific for the phosphorylated form of ERK1/2. C, a monoclonal anti-ERK2 antibody was used to detect the change of electrophoretic mobility of ERK2 in PGE2-stimulated HEL cells. These blots are representative of three independent experiments.

[³H]Thymidine Uptake. Proliferation of HEL cells was determined by incorporation of [³H]thymidine. Briefly, cells (1-2 × 10⁵ cells/ml) were incubated overnight in six well plates in serum-free RPMI 1640. Cells were then stimulated with various agents at the concentrations indicated and incubated in the presence of 1 µCi/ml [³H]thymidine. Twenty-four hours later, cells were washed three times with RPMI 1640, and the radioactivity associated with cells was analyzed by liquid scintillation counting. More than 95% of cells were still viable after 2 days of starvation, as assessed by trypan blue exclusion.

Data Analysis. The intensity of the ERK1/2 phosphorylation was scanned, quantitated by densitometry, and analyzed by Pharmacological Calculation System software (version 4.2 introduced in the 1980s; Springer-Verlag, New York, NY). Data in Table 1 are expressed as -fold of basal, defined as phosphorylation level of ERK1/2 in stimulated cells relative to that of nonstimulated, and analyzed by t test. In Figs. 2 and 3, the stimulatory response in the absence of inhibitors is normalized as 100%, and effects of various inhibitors are expressed as percentage of the response of the agonist. For each inhibitor, ERK1/2 phosphorylation induced by one specific activator in the absence or presence of inhibitor was compared, and the effects of PGE₂ and thrombin in cells treated with the same inhibitor were further analyzed, using ANOVA and the Newman-Keuls multiple comparison test. The effect of various treatments on [³H]thymidine uptake (Table 2) was also analyzed by ANOVA and the Newman-Keuls test. In all cases, P < 0.05 is considered to be statistically significant.

View larger version (44K): [in this window] [in a new window]   Fig. 2.   Effect of pertussis toxin on ERK1/2 activation. HEL cells were incubated overnight in serum-free medium in the presence or absence of pertussis toxin (100 ng/ml), followed by stimulation for 5 min with PGE2 (1 µM) or thrombin (5 U/ml) or for 10 min with PMA (100 nM). A typical blot probed with a phospho-specific ERK1/2 antibody is shown in panel A. The effect of agonist-induced ERK1/2 phosphorylation in the absence of pertussis toxin is set as 100%, and the inhibitory effect of pertussis toxin is expressed as percentage of the response of agonist (panel B). Data are mean ± S.E. of three to five experiments. CTL, control; PTX, pertussis toxin; Thr, thrombin. The effects of pertussis toxin on thrombin and PGE2 responses were analyzed by ANOVA and Newman-Keuls test. **, significantly different from PGE2 response in the absence of pertussis toxin, P < 0.01. , significantly different from the response of thrombin treated with pertussis toxin, P < 0.05.

View larger version (52K): [in this window] [in a new window]   Fig. 3.   ERK1/2 activation induced by PGE2 and thrombin shows differential sensitivity to various inhibitors. HEL cells were pretreated with or without inhibitors for appropriate times and then stimulated with PGE2 (1 µM), thrombin (5 U/ml), or PMA (100 nM). Times of pretreatment for various inhibitors were 30 min for BAPTA/AM (40 µM), 60 min for bisindolylmaleimide I (Bis I, 5 µM), 30 min for genistein (100 µM), 16 h for herbimycin (1 µM), 30 min for wortmannin (200 nM), and 10 min for U0126 (20 µM). The effects of inhibitors on ERK1/2 phosphorylation are normalized and expressed as a percentage of the response obtained from cells stimulated in the absence of inhibitors (panels A-C). The effects of inhibitors on basal ERK1/2 phosphorylation are summarized in panel D. Data are mean ± S.E. from three to five experiments. Multiple comparison analysis was performed by ANOVA and Newman-Keuls test. *, significantly different from the response of agonist in the absence of inhibitors, P < 0.05. **, significantly different from the response of agonist in the absence of inhibitors, P < 0.01. , significantly different from the response of thrombin with the same pretreatment, P < 0.05.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Expression and Phosphorylation of ERK1/2 in HEL Cells. To determine the occurrence of ERK in HEL cells, total lysates were separated by gel electrophoresis and analyzed by Western blotting using a polyclonal antibody that recognizes ERK1/2 and a monoclonal antibody that is specific for ERK2. As shown in lane 1 of Fig. 1A, the polyclonal antibody detected two bands with molecular mass approximately at 42 and 44 kDa. The major band (42 kDa) is probably ERK2 because the monoclonal anti-ERK2 antibody also recognized a protein migrating at the same location (Fig. 1A, lane 2). These results demonstrate that both ERK1 and ERK2 are expressed and suggest that the latter is likely to be the predominant form in HEL cells.

ERK1/2 Activation Is Receptor-Specific. In addition to the PGE₂-interacting EP receptors, a variety of G protein-coupled receptors have been characterized in HEL cells, including thrombin, ₂, P_2y, and A_2b receptors (Michel et al., 1989; Brass et al., 1991; Feoktistov et al., 1994; Baltensperger and Porzig, 1997). To determine whether these receptors can induce ERK1/2 activation, cells were stimulated for 5 min with adenosine, nucleotides, thrombin, or epinephrine. ERK1/2 were found to be activated strongly in response to thrombin and epinephrine (Table 1). Table 1 also shows that the response to serum approached that of PGE₂. In contrast, ADP, ATP, UTP, and adenosine exhibited little if any effect on ERK1/2 activity at concentrations up to 100 µM. These observations indicate that ERK1/2 are selectively coupled to G protein-coupled receptors.

ERK1/2 Activation Is Not Coupled to Proliferation. After having measured the effects of G protein-coupled receptor agonists, we continued to investigate the functional consequence of ERK1/2 activation. Because activation of ERK1/2 and proliferation are correlated after the stimulation of mitogen in many types of cells (Widmann et al., 1999), our first effort was to test whether ERK1/2 activation by G protein-coupled receptors is associated with proliferation in HEL cells. To test this possibility, thymidine uptake was measured in cells treated with various agonists. Surprisingly, strong ERK1/2 activators, such as PGE₂ and thrombin, were ineffective at inducing proliferation (Table 2). Of the agonists tested, ADP is the only agent that was able to evoke an increase of thymidine uptake to a level close to that of serum. Thus, these results indicate that ERK1/2 activation by G protein-coupled receptors is not necessarily associated with mitogenesis in HEL cells. In addition, it is interesting to note that ATP, although without effect by itself, blocked the response of ADP. This antagonistic phenomenon was also observed on ADP-induced calcium mobilization in HEL cells (Shi et al., 1995).

Effect of Pertussis Toxin on ERK1/2 Activation. To continue to explore the signaling pathways linking G protein-coupled receptors to ERK1/2 activation, thrombin and PGE₂ were used in subsequent experiments because, among the agonists tested, these two agonists produced the greatest effects (Table 1). To begin with, we investigated whether ERK1/2 phosphorylation in HEL cells is sensitive to pertussis toxin, as previous studies have shown that receptors of thrombin and PGE₂ in HEL cells transduce signals via multiple types of G proteins (Brass et al., 1991; Wu et al., 1991, 1992; Brass and Woolkalis, 1992). At 100 ng/ml, pertussis toxin significantly lowered basal ERK1/2 phosphorylation to 47 ± 23% of control (n = 5, P < 0.05 versus basal). Under the same condition, pertussis toxin nearly abolished ERK1/2 activation in response to PGE₂ (n = 5, P < 0.01, Fig. 2). This indicates that the EP receptors in HEL cells transduce their signals via the pertussis toxin-sensitive G proteins, probably the G_i family, because G_o is undetectable in HEL cells (Michel et al., 1989). On the other hand, toxin treatment had little inhibitory effect on ERK1/2 activation by thrombin or PMA. An increase of [Ca²⁺]_i induced by thrombin in HEL cells was previously demonstrated to be largely insensitive to pertussis toxin as well (Schwaner et al., 1992), suggesting that the effect of thrombin can be attributed to the G_q/11 family.

Agonist-Induced ERK1/2 Activation is Mimicked by Ca²⁺ Ionophore and Protein Kinase C Activator. We next asked which signaling pathways downstream of G proteins are involved in ERK1/2 activation. Previous studies have shown that E-series prostaglandins and thrombin activate adenylyl cyclase and phospholipase C in HEL cells (Brass et al., 1991; Wu et al., 1991; Brass and Woolkalis, 1992). We reasoned that if these pathways are involved, agents that directly stimulate the production or mimic the actions of second messengers should activate ERK1/2 as well. As expected, ionomycin and PMA stimulated ERK1/2 phosphorylation to a level similar to those of G protein-coupled receptor agonists (Table 1), implying that G protein-coupled receptors elicit their effects in part through phosphoinositide hydrolysis. The possibility of an involvement of cAMP is argued against by the findings that pertussis toxin inhibits PGE₂-induced ERK1/2 activation (Fig. 2) and that adenosine stimulates cAMP formation (Feoktistov and Biaggioni, 1993) but not ERK1/2 activity (Table 1). Two approaches were used to address the role of cAMP directly. ERK1/2 activity was examined in cells treated with forskolin, which stimulates adenylyl cyclase. Although forskolin induced a substantial increase of ERK1/2 phosphorylation, the magnitude was smaller than with thrombin or PGE₂ (Table 1). Activation of ERK1/2 was also weak with the permeable cAMP analog, 8-bromo-cAMP (Table 1). Although a modulatory role cannot be excluded, these findings suggest that the G_s-adenylyl cyclase pathway is not the primary mechanism leading to ERK1/2 activation in HEL cells.

Ca²⁺ Dependence of ERK1/2 Activation. If Ca²⁺ mediates the effect of agonist, it would be expected that ERK1/2 phosphorylation in response to thrombin and PGE₂ would be diminished by approaches that block calcium mobilization. To further address this issue, ERK1/2 phosphorylation was measured in cells that were loaded with BAPTA/AM. The cell-permeable BAPTA/AM has been shown to block Ca²⁺ mobilization by PGE₂ in HEL cells (Wu et al., 1992). Figure 3, A and B, shows that BAPTA/AM reduced ERK1/2 phosphorylation in response to PGE₂ by 83% (n = 5, P < 0.01) and to thrombin by 43% (n = 3, P < 0.01). The extent of inhibition by BAPTA is significantly different between PGE₂ and thrombin (P < 0.05). On the other hand, the basal phosphorylation and the response to PMA were not significantly affected (Fig. 3, C and D), indicating that ERK1/2 are not directly inhibited by BAPTA. These findings suggest that Ca²⁺ plays a key role in regulating ERK1/2 activity by PGE₂ and thrombin.

An Involvement of Protein Kinase C in ERK1/2 Activation. As described above, a role of protein kinase C is supported by the finding that protein kinase C activator PMA stimulated ERK1/2 phosphorylation (Table 1; Fig. 2). To further explore this possibility, ERK1/2 phosphorylation was measured in cells treated with or without bisindolylmaleimide I, a protein kinase C inhibitor. Figure 3D shows that this inhibitor had no significant effect on basal phosphorylation of ERK1/2. Under the same condition, PMA was unable to cause ERK1/2 activation (Fig. 3C), and the effects of PGE₂ and thrombin were lowered to 66 ± 6% and 25 ± 8% of control (n = 5, P < 0.01), respectively (Fig. 3, A and B). These results not only indicate a differential contribution of protein kinase C in the process linking receptors to ERK1/2 but also imply that other mechanisms may be involved as well.

Role of Tyrosine Kinases in ERK1/2 Activation. Previous studies have documented that both the G_i- and G_q/11-coupled receptors could transduce signals to ERK1/2 via tyrosine kinase-dependent pathways (Hawes et al., 1995; Wan et al., 1997). To further understand how ERK1/2 are regulated, we tested the effect of tyrosine kinase inhibitors in HEL cells. Although a small portion of the PGE₂ effect was inhibited by genistein, this is not significant (n = 5, Fig. 3A). To further examine the role of tyrosine kinases, we used herbimycin, a more potent and selective inhibitor. As shown in Fig. 3A, a significant portion of the PGE₂ effect was blocked by herbimycin (n = 3, P < 0.01). Figure 3 also shows that neither genistein nor herbimycin had effects on ERK1/2 phosphorylation induced by thrombin or PMA, suggesting that the pertussis toxin-insensitive response of thrombin is independent of tyrosine kinases, whereas the G_i-coupled EP receptors could mediate its effect via tyrosine kinases.

Role of PI 3-Kinase in ERK1/2 Activation. PI 3-kinase generates second messengers through phosphorylation of phosphoinositides at the D-3 position and transduces signals via protein kinase C, p70 S6 kinase, and Akt kinase (Toker and Cantley, 1997). Recently, participation of PI 3-kinase in linking G protein-coupled receptors to ERK1/2 activation has been demonstrated (Lopez-Illasaca et al., 1997; Keffel et al., 1999). We therefore were interested in determining whether PI 3-kinase is also involved in ERK1/2 activation in HEL cells. To address this possibility, wortmannin, a PI 3-kinase inhibitor, was used. Figure 3 shows that wortmannin had no significant effect on basal or PGE₂- and PMA-induced ERK1/2 phosphorylation, but it reduced the response of thrombin to 51 ± 10% of control (n = 5, P < 0.05), implying an involvement of PI 3-kinase in ERK1/2 activation by thrombin receptors.

ERK1/2 Activation is Dependent on MEK. A final series of experiments was designed to determine whether ERK1/2 are regulated by MAPK kinase (MEK). Since this regulatory pathway has been demonstrated in many types of cells (Widmann et al., 1999), we anticipated that ERK1/2 activation by thrombin and PGE₂ is regulated by MEK as well. In this regard, we used MEK inhibitor U0126 to test this possibility. Stimulation of ERK1/2, but not JNK and p38 MAPK, was shown to be selectively blocked by U0126 (Favata et al., 1998). As expected, neither PGE₂ nor thrombin stimulated ERK1/2 phosphorylation in cells treated with U0126 (Fig. 3, A and B). In addition, the effect of PMA was also abolished by this inhibitor (Fig. 3C). Under the same condition, U0126 reduced basal ERK1/2 phosphorylation to 34 ± 9% of control (Fig. 3D) but had no effect on cell viability or on phosphorylation of p38 MAPK induced by sorbitol (data not shown). These results suggest that, although thrombin and EP receptors mediate effects through distinctive pathways, they converge at the level of MEK to regulate ERK1/2 activity in HEL cells.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

In the present study, we have examined the relationship between G protein-coupled receptors and ERK1/2 activation in megakaryocytic cells. Like many types of cells, ERK1 and ERK2 are expressed in HEL cells and become phosphorylated in response to extracellular stimuli. Interestingly, ERK1/2 in these cells are selectively activated by certain agonists, including thrombin, epinephrine, and E-series prostaglandins (Table 1), as well as neuropeptide Y (Keffel et al., 1999). One possible explanation for this phenomenon is that only those receptors coupled to specific types of G proteins are able to activate ERK1/2. Based on data obtained from this and previous studies, it is clear that the G_i-coupled EP and neuropeptide Y receptors and the G_q/11-coupled thrombin receptor could mediate activation of ERK1/2 in these cells, whereas the G_s-coupled A_2b receptor does not transduce signals to ERK1/2. Given the evidence that none of the nucleotides tested in this study induce ERK1/2 phosphorylation (Table 1) and ATP and UTP mobilize [Ca²⁺]_i via G₁₆ in HEL cells (Baltensperger and Porzig, 1997), receptors coupled to G₁₆ are ineffective in activating ERK1/2. Collectively, these observations lead us to postulate that receptors coupled to either G_i or G_q/11 can be functionally linked to ERK1/2 in HEL cells.

Even though most of the agonists tested in HEL cells cause an increase of [Ca²⁺]_i (Motulsky and Michel, 1988; Michel et al., 1989; Brass et al., 1991; Schwaner et al., 1992; Wu et al., 1992; Feoktistov et al., 1994), our results provide evidence that not all the Ca²⁺-mobilizing receptors are capable of transducing signals to ERK1/2. These data suggest that an increase of [Ca²⁺]_i alone is not sufficient to activate ERK1/2, and other signaling pathways are involved as well. Elucidation of these pathways will gain more insight on how the ERK1/2 signaling specificity is accomplished, but it is not clear which pathways can be accounted for. We speculate that the answers may lie in the early steps of transmembrane signaling, perhaps at the level of G proteins and/or receptors, because they may transduce signals through additional pathways other than calcium mobilization. These additional pathways may amplify the effect of Ca²⁺ or recruit signaling components into a specific compartment so that ERK1/2 can be activated more efficiently. Given the diversity of the receptors and G proteins, it is conceivable that specific pathways may be activated by certain Ca²⁺-mobilizing receptors but not by others, and therefore ERK1/2 can be activated in a receptor-specific manner.

In contrast to earlier studies in HEL cells, there are differences regarding the activation and regulatory pathways of ERK1/2 by EP and thrombin receptors. First, unlike previous results reported by Keffel et al. (1999), this study showed that thrombin stimulates ERK1/2 phosphorylation in HEL cells. The reasons for this discrepancy are not clear, but it is without precedent. Brass and Woolkalis (1992) found that thrombin stimulates cAMP production, whereas data from Turner et al. (1992) indicated that it has no effect on adenylyl cyclase activity. As HEL cells are pluripotent cells, it is possible that this cell line may spontaneously differentiate in culture, causing cells to respond to thrombin in a different way. Second, although previous studies in HEL cells have shown that ERK1/2 activation by neuropeptide Y receptor is not mediated by protein kinase C (Keffel et al., 1999), protein kinase C inhibitor attenuates effects of PGE₂ and thrombin (Fig. 3). These findings suggest that an involvement of protein kinase C in ERK1/2 activation may be receptor-specific. Third, our results reveal that part of the effect of G_q/11-coupled thrombin receptor, but not the G_i-coupled EP receptors, is sensitive to wortmannin (Fig. 3). This is in contrast to the notion that G_i-mediated ERK1/2 activation involves PI 3-kinase (Lopez-Illasaca et al., 1997; Keffel et al., 1999). The difference of wortmannin susceptibility may be accounted for by two reasons. One possibility is that thrombin receptor is capable of interacting with both G_i and G_q/11, as evidenced by the findings that the effects of thrombin in HEL cells are partially sensitive to pertussis toxin (Brass et al., 1991; Brass and Woolkalis, 1992). Alternatively, HEL cells may express a high level of EP receptors, and these receptors induce ERK1/2 activation through multiple pathways. As a result, this effect would not be affected by wortmannin even though PI 3-kinase is inhibited because other mechanisms present in these cells may take the place of PI 3-kinase.

Our observation that ERK1/2 are activated by G protein-coupled receptors in HEL cells suggests a role of these kinases in cellular functions. On the basis of its importance in proliferation, it is expected that ERK1/2 activation may be associated with mitogenesis in HEL cells. However, we have found that strong ERK1/2 activators did not stimulate thymidine uptake of these cells (Table 2), implying that ERK1/2 may not be an important mediator in linking G protein-coupled receptors to proliferation. Evidence from receptor tyrosine kinases suggests that transient activation of ERK1/2 induces proliferation (Marshall, 1995), but results presented in this study show that transient activation is not associated with an increase of DNA synthesis. It is possible that G protein-coupled receptors may require additional signaling pathways to initiate the cell cycle machinery or that ERK1/2 engage in other cellular functions in HEL cells. Most of the studies in megakaryocytes or K562 erythroleukemia cells indicate that ERK1/2 are involved in differentiation (Racke et al., 1997; Whalen et al., 1997; Fichelson et al., 1999; Rojnuckarin et al., 1999). Perhaps the outcome of ERK1/2 activation by G protein-coupled receptors in HEL cells is to regulate maturation rather than growth. Experiments are under way to elucidate the downstream events and the functional significance of ERK1/2 activation in HEL cells.

Another interesting finding is that ADP stimulated thymidine uptake in HEL cells. ADP has also been shown to stimulate proliferation in aortic smooth muscle cells (Wang et al., 1992). However, ATP is mitogenic in smooth muscle cells, whereas it is without effect in this study. These findings imply that the receptor with which ADP interacts in HEL cells is different from that in vascular smooth muscle. Based on the antagonistic effect of ATP on thymidine uptake (Table 2) and calcium mobilization (Shi et al., 1995), it would suggest that the ADP receptor in HEL cells is similar to those in platelets (i.e., P_2Y1 and P_2Y12 receptors) (Kunapuli and Daniel, 1998; Hollopeter et al., 2001; Zhang et al., 2001). Since ADP does not inhibit adenylyl cyclase in HEL cells (Vittet et al., 1992), the P_2Y1 receptor cloned from these cells (Ayyanathan et al., 1996) may be responsible for the ADP effect observed in this study. In addition to the receptor identity, the G proteins and the signaling mechanisms linking the ADP receptor to proliferation are not known in this megakaryocytic cell line. Earlier studies showed that ADP induces calcium mobilization (Schwaner et al., 1992; Shi et al., 1995), and its effect is not influenced in G₁₆-deficient HEL cells (Baltensperger and Porzig, 1997). Our result further indicates that ADP has no significant effect on ERK1/2 activation. More studies are required to identify the G proteins and the signaling pathways coupled to ADP receptor.

The information obtained from the present study not only has advanced our understanding of signaling pathways of PGE₂ but also raises the question with respect to the receptors accounting for ERK1/2 activation. In contrast to its stimulatory effect in HEL cells, studies in renal mesangial and airway smooth muscle cells indicated that PGE₂ is inhibitory to ERK1/2 (Li et al., 1995; Lee et al., 2001). The opposite effects of PGE₂ may be attributed to the receptors expressed in these cells. Recent cloning has demonstrated that the EP₁ and EP₃ subtypes are coupled to phosphoinositide turnover, calcium mobilization, and inhibition of adenylyl cyclase, whereas the EP₂ and EP₄ subtypes stimulate cAMP production (Coleman et al., 1994). Since PGE₂-induced ERK1/2 activation in HEL cells is dependent on pertussis toxin-sensitive G proteins and an increase of [Ca²⁺]_i, it is likely that the EP₁ and EP₃ receptors mediate the effect of PGE₂. This possibility is supported by the findings that these two receptors have been cloned from HEL cells (Funk et al., 1993; Kunapuli et al., 1994). Further studies using subtype-selective ligands and transfection system may help to determine the identity of receptors with which E-series prostaglandins interact.