Abstract
Mitogenic signaling of G protein-coupled receptors (GPCRs) can proceed via sequential epidermal growth factor receptor (EGFR) transactivation and extracellular signal-regulated kinase (ERK) phosphorylation. Although the μ-opioid receptor (MOR) mediates stimulation of ERK via EGFR transactivation in human embryonic kidney 293 cells, the mechanism of acute MOR signaling to ERK has not been characterized in rat C6 glioma cells that seem to contain little EGFR. Herein, we describe experiments that implicate fibroblast growth factor (FGF) receptor (FGFR) transactivation in the convergence of MOR and growth factor signaling pathways in C6 cells. MOR agonists, endomorphin-1 and morphine, induced a rapid (3-min) increase of ERK phosphorylation that was abolished by MOR antagonistd-Phe-Cys-Tyr-d-Trp-Arg-Thr-Pen-Thr-NH2. By using selective inhibitors and overexpression of dominant negative mutants, data were obtained to suggest that MOR signaling to ERK is transduced by Gβγ and entails Ca2+- and protein kinase C-mediated steps, whereas the FGFR branch of the pathway is Ras-dependent. An intermediary role of FGFR1 transactivation was suggested by MOR- but not κ-opioid receptor (KOR)-induced FGFR1 tyrosine phosphorylation. A dominant negative mutant of FGFR1 attenuated MOR- but not KOR-induced ERK phosphorylation. Thus, a novel transactivation mechanism entailing secreted endogenous FGF may link the GPCR and growth factor pathways involved in MOR activation of ERK in C6 cells.
One of the most interesting examples of cross talk between cellular signaling systems is the inter-relationship between GPCR and receptor tyrosine kinase (RTK) pathways leading to mitogen-activated protein kinase activation. Due in part to the diversity of the GPCR superfamily of proteins, several mechanisms of convergence of the two pathways have been detected. Recent findings reveal that one mechanism of heterologous GPCR signaling to ERK occurs via tyrosine phosphorylation of the RTK itself (Daub et al., 1996; Prenzel et al., 1999). Transactivation of EGFR rapidly ensues upon agonist stimulation of a broad range of GPCRs, including MOR. This mechanism is cell type-specific, because lysophosphatidic acid (LPA) stimulation of ERK is EGFR transactivation-dependent in Rat-1 cells but not in PC12 cells (Della Rocca et al., 1999). In HEK293 cells, MOR, LPA, and β2-adrenergic receptor-mediated ERK activation is partially dependent on EGFR transactivation and minor alternative pathways exist (Della Rocca et al., 1997; Belcheva et al., 2001). EGFR transactivation may entail a mechanism wherein plasma membrane-bound MMPs shed EGF-like precursor molecules anchored on the cell surface (Prenzel et al., 1999).
Evidence for GPCR transactivation of other RTKs has been reported. These include angiotensin-induced PDGFR tyrosine phosphorylation in vascular smooth muscle cells (Linseman et al., 1995), LPA-stimulated PDGFR tyrosine phosphorylation in L cells (Herrlich et al., 1998), and dopamine D-4 and D-2L receptor transactivation of PDGFR in Chinese hamster ovary cells (Oak et al., 2001). In these pathways, blockade of RTK tyrosine phosphorylation results in attenuation of GPCR-mediated ERK phosphorylation. The mechanism of PDGFR transactivation is not as well characterized as that of EGFR. Information on transactivation of other RTKs is sparse.
C6 cells are known to express a number of growth factors, including FGF, IGF-1, PDGF, and vascular endothelial growth factor in addition to their receptors (Okumura et al., 1989; Chernausek, 1993; Hamel and Westphal, 2000). FGF may stimulate cell growth, angiogenesis, and proliferation during development, wound healing, and in neoplasia (Powers et al., 2000). bFGF is broadly expressed in neurons and glia of the central nervous system where, in addition to its mitogenic properties, it elicits neuroprotective effects (Kinkl et al., 2001) and promotes process outgrowth in oligodendrocytes during myelination (Oh et al., 1999).
Recently, we examined signal mechanisms mediated by endogenous KOR in C6 cells (Bohn et al., 2000a). We found that the KOR agonist U69,593 stimulated phospholipase C, ERK phosphorylation, PYK2 phosphorylation, and DNA synthesis. U69,593-stimulated ERK activation was shown to be upstream of DNA synthesis because inhibition of pertussis toxin (PTX)-sensitive G proteins, L-type Ca2+ channels, phospholipase C, intracellular Ca2+ release, protein kinase C (PKC), and ERK kinase blocked both ERK activation and DNA synthesis. We also obtained evidence to suggest that ERK activation is Ras-dependent and transduced by Gβγ subunits. A schematic presentation of the intermediates involved in KOR signaling to ERK has been published previously (Belcheva and Coscia, 2002).
In addition to KOR, C6 cells express MOR that also modulates DNA synthesis (Barg et al., 1994; Bohn et al., 2000b). Although chronic μ-opioids were shown to inhibit KOR- mediated and other mitogen receptor-mediated stimulation of ERK in C6 cells, the mechanism of acute μ-opioid signaling to ERK has not been investigated. In this study, we implicate Gβγ, Ras, Ca2+, PKC, metalloprotease, and FGFR in MOR-induced ERK phosphorylation in C6 cells. Because MOR promotes ERK phosphorylation by EGFR transactivation in HEK293 cells (Belcheva et al., 2001), it was of interest to characterize the mechanism of acute MOR signaling to ERK and the point of convergence with RTK pathways in C6 cells that seem to contain little EGFR. The evidence gained herein suggests a novel transactivation mechanism wherein FGFR is at the site of convergence between growth factor pathways and MOR signaling to ERK in C6 cells.
Materials and Methods
Reagents.
Chemicals were purchased from Sigma-Aldrich (St. Louis, MO) with the following exceptions: [d-Ala2,N-Me-Phe4,Gly5-ol]-enkephalin (DAMGO), [3H]DAMGO, and CTAP were obtained from Multiple Peptide Systems (San Diego, CA); U69,593 was from National Institute on Drug Abuse Drug Supply (Research Triangle, NC); FGF (human, recombinant), EGF (human, recombinant), Dulbecco's modified Eagle's medium (DMEM), minimal essential media (MEM), calf serum (CS), Lipofectin, and LipofectAMINE were from Invitrogen (Carlsbad, CA); FuGENE 6-transfection reagent was from Roche Applied Science (Indianapolis, IN); fetal bovine serum was from Harlan Bioproducts for Science (Indianapolis, IN); tyrphostin AG1478, phorbol 12-myristate 13-acetate (PMA), BAPTA, and bisindolylmaleimide I (GFX) were from Calbiochem (San Diego, CA); SU5416 and SU6668 were from Sugen (San Francisco, CA); anti-phospho-ERK antibody and anti-phospho-tyrosine antibody (P-Tyr-100, mouse monoclonal) were from Cell Signaling Technology (Beverly, MA); anti-ERK antibody and anti-FGFR1 antibody (rabbit polyclonal) were from Santa Cruz Biotechnology (Santa Cruz, CA); and protein G-agarose suspension was from Oncogene Science (Cambridge, MA).
Cell Cultures.
C6 cells (American Type Culture Collection, Manassas, VA) were maintained under conditions in which levels of glucose, glutamine, inositol, and serum were controlled in the growth media as described previously (Bohn et al., 2000a). Briefly, cells were initially grown in DMEM + 10% fetal bovine serum (heat-inactivated) for two passages. Media were then replaced with DMEM + 5% CS (heat-inactivated) and cells were maintained for an additional 8 to 10 passages. Cells were used between passages 50 to 62. In each experiment, superconfluent cells were collected in phosphate-buffered saline-EDTA and upon centrifugation, pellets were resuspended and plated in DMEM + 5% CS in six-well plates. After allowing cells to recover overnight and to adhere to the plate surface, optimal starvation was achieved with the following media: MEM lacking glucose, inositol, and glutamine (prepared by Washington University Tissue Culture Laboratories, St. Louis, MO) with 10% MEM to yield final concentrations of 100 mg/l glucose and 0.2 mg/l inositol (serum andl-glutamine free). Cells were maintained in this “low MEM” for 48 h before inhibitor or agonist treatment. In all assays, agonists, antagonists or inhibitors were delivered in low MEM.
Stable and Transient Transfections.
C6 cells were stably transfected with rat MOR cDNA (pCMV-neo expression vector) using Lipofectin according to the manufacturer's description. For transient transfections, C6 cells were plated in DMEM + 5% CS at about 200,000 cells/well in six-well plates. After overnight growth, wells were approximately 70% confluent. Cells were washed two times in MEM and were transfected with 2 μg of cDNA [CD8, CD8-β-adrenergic receptor kinase (CD8-βARK-C) or the dominant negative mutant N17-Ras] and LipofectAMINE, as described previously (Belcheva et al., 1998). pcDNA-CD8-βARK-C expresses the extracellular and transmembrane domains of CD8 fused to an intracellular domain containing the carboxyl terminus of β-adrenergic receptor kinase (the βγ binding portion). After overnight incubation, transfection media were replaced with DMEM + 5% CS, and cells were allowed to recover for 48 h. FuGENE 6 Transfection Reagent was used for transient transfections of the dominant negative mutant of FGFR1, or of pcDNA3 (for mock transfections) following the manufacturer's instructions and using 1 μg of cDNA and 3 μl of transfection reagent. After 48 h, media were replaced with low MEM. In parallel samples, overexpression was verified by immunoblot analysis with either anti-CD8 (CD8 was undetectable in untransfected cells), anti-Ras antibodies (yielded 10-fold greater Ras immunoreactivity than vector alone), or anti-FGFR1 antibodies (yielded 4-fold greater FGFR1 immunoreactivity than vector alone).
ERK Assays.
For most of the studies presented herein, ERK phosphorylation was monitored by immunoblotting (Bohn et al., 2000a;Belcheva et al., 2001). Briefly, this method includes the following steps. Antagonists and inhibitors were added to the media for the specified times before stimulation with agonist, respectively. Cell lysates were collected in lysing buffer (20 mM HEPES, 10 mM EGTA, 40 mM β-glycerophosphate, 2.5 mM MgCl2, 2 mM sodium vanadate, 1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 20 μg/ml aprotinin, and 20 μg/ml leupeptin, pH 7.5). Protein assays were performed using the Bradford reagent. Equivalent amounts of protein were loaded per lane (10–20 μg) on 10% SDSPA mini-gels. Western blots were performed using anti-phospho-ERK antibodies and peroxidase-conjugated mouse secondary antibody. Data are expressed as -fold over control ± inhibitor. In representative experiments, gels were stripped and reprobed with anti-ERK antibody to ascertain that equivalent amounts of ERK were present in each lane. Polyacrylamide gel electrophoresis bands were visualized by chemiluminescence and band intensities were determined by densitometric analysis using a Kodak DC120 digital (1.2 mega pixel) camera, Kodak ds 1D version 3.0.2 software (Scientific Imaging Systems, New Haven, CT), and Scion Image PC software for NIH Image version 1.62 (Scion Corporation, Frederick, MD).
In experiments using cells stably transfected with MOR, ERK activity was analyzed by the “in-gel” kinase method with modifications as described previously (Belcheva et al., 1998).
MOR Binding Experiments.
Cells stably transfected with MOR were harvested and homogenized by gentle disruption in a “cell cracker”, and membranes were assayed for binding as described previously (Belcheva et al., 1993). A membrane fraction (P20) was prepared from cell homogenates by sedimenting a 1000-g supernatant at 20,000g. A cocktail containing 10 μg/ml leupeptin, 2 μg/ml pepstatin A, 200 μg/ml bacitracin, and 1 mM phenylmethylsulfonyl fluoride was added to Tris buffer used for preparation of this membrane fraction. Membranes (300–500 μg/ml) were incubated with 1 nM [3H]DAMGO (35 Ci/mmol) at room temperature for 1 h. Nonspecific binding was determined in the presence of 1 to 10 μM DAMGO. Reactions were terminated by addition of cold Tris buffer to the tubes followed by rapid filtration over GF/B filters in a cell harvester (Brandel Inc., Gaithersburg, MD). Filters were washed twice with cold 50 mM Tris-HCl, pH 7.4, buffer and then counted. Binding parameters (K d andB max values) were determined by using the LIGAND program.
FGFR1 Immunoprecipitation.
Cells were serum-deprived for 24 h and treated with agonist. Cultures were lysed with a modified radioimmunoprecipitation assay buffer. FGFR1 was immunoprecipitated with a rabbit polyclonal anti-FGFR1 antibody and 20 μl of protein G-Sepharose beads per sample. After 7.5% SDS-polyacrylamide gel electrophoresis, proteins were blotted with p-Tyr antibody and peroxidase-conjugated mouse secondary antibody.
Statistical Analysis.
Statistical determinations were made by Student's t test analysis using GraphPad Prism software version 2.01 (GraphPad Software, San Diego, CA). All data are expressed as the mean ± S.E.M.
Results
Endogenous MOR Mediates a Rapid, Acute Stimulation of ERK Phosphorylation.
The time course of μ-opioid stimulation of ERK phosphorylation was compared with that of the mitogen bFGF in C6 cells. As shown in Fig. 1A, μ-opioids endomorphin-1 and morphine elicited an ephemeral stimulation of ERK that was maximal in 3 min. In contrast, exogenous bFGF induced a more potent and longer lasting effect than the opioids with optimal values at 5 to 30 min. MOR antagonist CTAP blocked endomorphin-1 and morphine stimulation of ERK phosphorylation (Fig. 1B). However, CTAP had no effect on basal ERK phosphorylation. The results indicate that endomorphin-1 and morphine stimulate ERK phosphorylation via endogenous MOR present in C6 cells.
Nifedipine, an Inhibitor of L-Type Calcium Channels, and Dantrolene, an Inhibitor of Intracellular Ca2+ Release, Abolish μ-Agonist Stimulation of ERK Phosphorylation.
To examine the mechanism involved in MOR stimulation of ERK phosphorylation, several selective inhibitors directed against potential signaling components of the pathway that were implicated in KOR signaling to ERK in C6 cells were tested. Inhibitors were used at concentrations close to their corresponding IC50 values in these and most of the following experiments. To investigate the role of Ca2+ in MOR activation of ERK, cells were preincubated with either dantrolene or nifedipine before endomorphin-1 or morphine addition. The data presented in Fig.2 indicate that both inhibitors abolished MOR stimulation of ERK phosphorylation. In the absence of agonist, the inhibitors had no effect on basal ERK levels as reported previously (Bohn et al., 2000a). The μ-stimulation of ERK phosphorylation seems to be dependent upon influx of Ca2+ via L-type calcium channels and intracellular stores in C6 cells, as seen for κ-signaling by Bohn et al. (2000a).
PKC Inhibitor GFX and PKC Down-Regulator PMA Attenuate the Stimulation of ERK Phosphorylation by the Endogenous MOR.
Because PKC is an important Ca2+ binding protein that has been frequently implicated in GPCR activation of mitogen-activated protein kinases, its possible involvement in MOR signaling was tested in C6 cells. Cells were preincubated with either PMA overnight or with GFX for 30 min before addition of either endomorphin-1 or morphine and ERK phosphorylation was measured by immunoblotting (Fig.3). GFX, a relatively selective inhibitor of PKC, abolished MOR-induced phosphorylation of ERK. Similar inhibition of endomorphin-1 signaling to ERK was observed upon down-regulation of PKC by PMA treatment. Although PMA displayed a trend of attenuation of morphine signaling, the effect was not statistically significant. As a positive control, acute PMA alone stimulates ERK phosphorylation in C6 cells (data not shown). The MOR pathway to ERK is PKC-dependent in C6 cells as seen for KOR signaling (Bohn et al., 2000a).
DAMGO Activation of ERK in MOR-Overexpressing C6 Cells Is Gβγ- and Ras-Dependent.
To assess the role of Gβγ and Ras in ERK signaling by MOR, we overexpressed interfering mutant proteins in C6 cells stably transfected with MOR. Binding parameters for the overexpressed μ-sites were K d = 2.0 ± 0.1 nM and B max = 1.0 ± 0.07 pmol/mg of protein for DAMGO, a μ-selective, potent agonist that has been used extensively to measure binding or signaling for overexpressed rat and human MOR in cells (Belcheva et al., 1998, 2001). Activation of ERK by DAMGO in the stably transfected C6 cells was ∼2-fold greater than in cells containing endogenous MOR (compare Fig.4 with Figs. 1-3). MOR-overexpressing cells were transiently transfected with plasmids containing either cDNA of RasN17, a dominant negative mutant of Ras or CD8-βARK-C, a membrane-anchoring protein (CD8) fused to the βγ subunit-binding, C-terminal segment of βARK. Cells were treated with either DAMGO or the mitogen endothelin and ERK activation was assayed. Expression of N17-Ras or CD8-βARK-C in the cells attenuated DAMGO activation of ERK, implicating Ras and Gβγ in μ-opioid regulation of ERK activity in C6 cells (Fig. 4A). Cotransfection of COS-7 cells with CD8 alone had no effect on opioid activation of ERK in previous studies (Belcheva et al., 1998). Stimulation of ERK activity by endothelin was also reduced by cotransfection with N17-Ras or CD8-βARK-C in the MOR-overexpressing C6 cells. The endothelin findings are consistent with previous data that showed that its mitogenicity is mediated via two pathways, one of which is PTX-sensitive (Gi/oprotein βγ subunits) and PTX-resistant (Gαq) pathways (Lin et al., 1992; Barg et al., 1994). An additional positive control was obtained by showing that EGF stimulation of ERK activity was reduced by cotransfection with N17-Ras in MOR- and KOR-overexpressing COS-7 cells (Belcheva et al., 1998). The results suggest that the mechanism of MOR stimulation of ERK activity in μ-overexpressing cells entails the intermediacy of Gβγ subunits and Ras.
DAMGO Activation of ERK in MOR-Overexpressing C6 Cells Is Wortmannin-Insensitive.
The involvement of PI3K in ERK activation has been suggested for several GPCRs in different cell lines (Belcheva and Coscia, 2002). Herein, C6 cells stably transfected with MOR were treated with different concentrations of wortmannin before addition of DAMGO and ERK activation was measured. The absence of inhibition of DAMGO-induced ERK activation in the presence of all concentrations of wortmannin tested suggests that PI3K is not involved in μ-agonist-mediated ERK activation in C6 cells (Fig. 4B). As a positive control, DAMGO stimulation of ERK phosphorylation was reduced by wortmannin in MOR-transfected COS-7 cells by 65% (P < 0.05).
Indolinone RTK Inhibitors Attenuate MOR Stimulation of ERK Phosphorylation.
Although we have implicated an EGFR transactivation step in MOR stimulation of ERK phosphorylation in HEK293 cells (Belcheva et al., 2001), we did not detect intact EGFR in C6 cells by immunoblotting (Fig. 5A). Instead, more FGFR1 was found in C6 cells and rat astrocytes than in HEK293 cells (Fig. 5B). EGF slightly increased ERK phosphorylation in C6 cell lysates but the change was not statistically significant (Fig.5C). In addition, tyrphostin AG1478, which is a specific inhibitor of EGFR Tyr kinase activity, had no effect on ERK modulation by either bFGF or the small change elicited by EGF, suggesting that EGFR is not the mediator of the actions of these two growth factors. Thus, involvement of other RTKs in C6 cells was investigated by using indolinone RTK tyrosine kinase inhibitors SU5416 or SU6668. SU5416 is a potent inhibitor of VEGFR, but it also affects FGFR and PDGFR at the concentrations used (Mendel et al., 2000). SU6668 is more selective for FGFR signaling, but it is an inhibitor of PDGFR and VEGFR activation as well (Laird et al., 2000). It was also shown that neither inhibitor interferes with EGFR signaling to ERK, even at doses as high as 100 μM. Both RTK inhibitors significantly attenuated ERK phosphorylation by endomorphin-1, morphine, and bFGF, suggesting that transactivation of an RTK may play a role in MOR signaling to ERK in C6 cells (Fig.6). SU5416 and SU6668 reduced MOR-induced stimulation of ERK to basal levels. Although some reduction in signaling to ERK by the KOR agonist U69,593 was observed, neither RTK inhibitor significantly affected U69,593-induced ERK phosphorylation. Unlike SU5416, the more selective inhibitor of FGFR signaling, SU6668, did not attenuate U69,593 stimulation of ERK phosphorylation because this effect was significantly higher than the control (Fig. 6). The inhibition of bFGF stimulation of ERK by both SU5416 and SU6668 served as a positive control, whereas AG1478 was a negative control.
MMP Inhibitors Abolish MOR and KOR Stimulation of ERK Phosphorylation.
The ability of C6 cells to express many RTKs (see Introduction), suggests potential paracrine and autocrine effects may occur in growth regulation in C6 cells. It has been well documented that MMPs are involved in the shedding of plasma membrane-bound EGF from its membrane anchor (Daub et al., 1996, 1997; Prenzel et al., 1999; Roudabush et al., 2000; Belcheva et al., 2001; Oak et al., 2001;Belcheva and Coscia, 2002). The use of MMP inhibitors may also provide information on putative transactivation of other RTKs such as PDGF or FGFR by their corresponding agonists that may be on the plasma membrane. A membrane-bound metalloendopeptidase that has been found to be associated with C6 membranes is inhibited byo-phenanthroline and is highly sensitive to phosphoramidon action (Amberger et al., 1994). Therefore, to study the possible role of growth factors and their receptors in MOR and KOR stimulation of ERK phosphorylation, cells were preincubated with eithero-phenanthroline or phosphoramidon before addition of endomorphin-1, morphine, and U69,593. Because MOR and KOR stimulation of ERK phosphorylation was significantly reduced by both MMP inhibitors, ectodomain shedding of growth factors leading to possible RTK transactivation may be involved in opioid signaling to ERK in C6 cells (Fig. 7).
Requirement of FGFR1 Transactivation in MOR, but not KOR, Stimulation of ERK Phosphorylation.
As discussed in the Introduction, the requirement of RTK transactivation in GPCR stimulation of ERK has been established predominantly for EGFR with few reports on other RTK involvement. Because immunoblot analysis and tyrphostin, AG 1478 results in Fig. 5C suggest that the EGFR is not present in C6 cells, studies on the possible requirement for FGFR1 transactivation step in MOR and KOR stimulation of ERK phosphorylation were undertaken.
To determine whether MOR mediates the phosphorylation of FGFR1, cells were treated with endomorphin-1, U69,593, or bFGF and FGFR1 tyrosine phosphorylation was measured by immunoprecipitation of lysates with FGFR1 antibody and immunoblotting with phosphotyrosine antibody (Fig.8A). Although endomorphin-1 and bFGF elicited a rapid stimulation of FGFR1 phosphorylation, U69,593 had no effect under the conditions used in these studies. When cells were pretreated with SU6668, it inhibited endomorphin-1-induced FGFR1 phosphorylation. In some experiments, phosphotyrosine immunoblots were stripped of antibody and probed with FGFR1 antibody, and gels were stained to ascertain that the appropriate bands in comparable amounts were measured for tyrosine phosphorylation (data not shown).
The above-mentioned data support the notion of a requirement for FGFR1 transactivation in MOR stimulation of ERK phosphorylation in C6 cells. However, other growth factors and their receptors are expressed in this line and the indolinones used in the studies were not only selective for FGFR alone but also inhibit PDGFR and VEGFR. Thus, we cannot rule out the possibility that MOR-induced FGFR1 activation occurs but does not lead to ERK phosphorylation. To further implicate FGFR, cells were transiently transfected with a dominant negative mutant of FGFR1 (FGFR1dn) and treated with opioids or bFGF. As seen in Fig. 8B, the presence of FGFR1dn reduced endomorphin-1 induction of ERK but failed to do so in the case of U69,593. bFGF-induced ERK phosphorylation was significantly inhibited by FGFR1dn (data not shown). The data support the view that FGFR1 is directly involved in MOR signaling to ERK, but this RTK may not participate in KOR signaling to ERK.
Discussion
The results presented herein suggest that acute MOR and KOR signaling to ERK share certain features, but they seem to differ in the sites of convergence of the GPCR and RTK branches of these two heterologous pathways in C6 cells. The activation of ERK by MOR was shown to be transduced by Gβγ subunits and involve L-type Ca2+ channels, intracellular Ca2+ release, and PKC as shown previously for KOR. Both the MOR and KOR pathways are Ras-dependent, consistent with previous findings on astroglia (for review, see Belcheva and Coscia, 2002). Evidence implicating the unprecedented FGFR1 transactivation as the convergent step was obtained for the MOR pathway. KOR did not induce FGFR1 tyrosine phosphorylation and only MOR stimulation of ERK phosphorylation was significantly attenuated by indolinones that are known to interfere with the tyrosine kinase activity of FGFR (Laird et al., 2000; Mendel et al., 2000). The finding that FGFR1dn attenuated MOR, but not KOR, stimulation of ERK in C6 cells, further supported this hypothesis.
MOR signaling to ERK in C6 cells was found to be Gi/oβγ-, Ca2+-, and PKC-dependent by using selective inhibitors of these signaling components. These findings are consistent with some of our previous data on MOR-mediated, chronic morphine inhibition of intracellular Ca2+ mobilization, phosphoinositol turnover, and DNA synthesis in C6 cells (Barg et al., 1994; Bohn et al., 2000b). Several signaling components have been implicated in opioid heterologous pathways to ERK in a number of cell types (Fukuda et al., 1996; Belcheva et al., 1998; Hedin et al., 1999; Schmidt et al., 2000), including those in C6 cells by Bohn et al. (2000a) and here. Some of the same signaling components (PKC, MMPs, and RTKs) identified as mediating MOR activation of ERK in C6 cells herein have also been implicated in MOR signaling to ERK in rat primary astrocytes (unpublished observations). Recently, multiple roles of calmodulin were found in both GPCR and RTK parts of MOR signaling to ERK in HEK293 cells (Belcheva et al., 2001). One calmodulin-requiring step depended upon the direct interaction of this Ca2+-binding protein with MOR. There are also reports of PKC-dependent and -independent MOR and nociceptin receptor signaling to ERK that may coexist in the same cells (Hawes et al., 1998; Belcheva et al., 2001).
On the basis of the data herein and previous findings, PI3K involvement seems to be cell type-specific (Daub et al., 1997; Duckworth and Cantley, 1997; Belcheva et al., 1998; Hawes et al., 1998; Ai et al., 1999). Interestingly, ERK phosphorylation was only sensitive to wortmannin at low levels of EGFR in COS-7 cells and at low levels of PDGF receptor in Swiss 3T3 cells (Daub et al., 1997; Duckworth and Cantley, 1997; Belcheva et al., 1998). These and other results also suggest the existence of a PI3K-dependent, redundant pathway to ERK when larger quantities of growth factor receptor molecules are potentiated. Thus, the lack of an effect with wortmannin in the absence of EGFR in C6 cells is consistent with previous findings.
There are reports suggesting that cAMP/protein kinase A signaling does not elicit MOR-mediated ERK phosphorylation, but in some cases it is inhibitory, implying the existence of cross talk between this system and signaling to ERK as well (Ai et al., 1999; Kramer et al., 2000). In primary neuronal cells or neuronal model systems such as PC12 cells, cAMP induces ERK phosphorylation via the small GTPase Rap1, which is activated by protein kinase A and interacts with B-Raf (Vossler et al., 1997). In contrast, in astrocytes and astrocytoma cells, which do not possess B-Raf, cAMP inhibits ERK phosphorylation by the classical Ras/cRaf-1 pathway (Dugan et al., 1999). Accordingly, if B-Raf is transfected into astrocytoma cells, cAMP stimulates ERK phosphorylation as it does in B-Raf-containing neuronal cells. Without B-Raf in astrocytes, the Ras pathway to ERK seems to play a major role in astrocytes. As discussed above, opioid heterologous signaling to ERK often entails a Ras-dependent RTK pathway. Because we grow C6 cells under conditions in which they express an astrocytic phenotype, the implication of Ras in the present studies is consistent with previous findings of the role of this G protein in mitogenesis of astroglial cells. It has been recognized for some time that opioids can exert their neurotrophic actions on “flat” (type 1) astrocytes in brain, in primary cultures, and in astrocytic model cells such as C6 cells (Stiene-Martin and Hauser 1990; Barg et al., 1993, 1994). The present study of the mechanism of opioid mitogenic action gains significance in light of recent evidence of the important role that astrocytes play in the formation, maintenance, and function of neuronal synapses in both developing and mature brain (Oliet et al., 2001; Ullian et al., 2001).
As discussed in the Introduction, there is ample evidence to implicate EGFR transactivation in GPCR (including MOR) heterologous signaling to ERK. Several instances of GPCR-mediated PDGFR phosphorylation have also been reported. The mechanism of EGFR transactivation is thought to entail a complex series of events wherein GPCR-induced MMP cleaves soluble EGF-like ligands (such as heparin-binding EGF) from their plasma membrane-bound anchoring domains, followed by binding of the ligands to EGFR (for review, see Pierce et al., 2001).
A recent report showed that IGF-1 activation mediated the transactivation of EGFR and thereby ERK phosphorylation (Roudabush et al., 2000). IGF-1 is expressed in astrocytes and C6 cells and is also thought to have autocrine/paracrine mitogenic actions in these cells (Chernausek, 1993). However, the lack of inhibition of EGFR-mediated ERK phosphorylation by the selective EGFR tyrosine kinase inhibitor AG1478 as well as the absence of EGFR in immunoblots of C6 cell lysates reduces the possibility of FGFR1 cross talk with EGFR in the activation of ERK that is comparable with that between IGF-1 and EGF.
The results shown herein provide the first evidence for FGFR1 transactivation in GPCR-mediated ERK activation. Our findings extend our previous mechanistic studies by demonstrating that the transactivation of FGFR1 is at the site of convergence between MOR and RTK signaling in C6 cells. FGF is synthesized in C6 cells and can be secreted to elicit proliferation via autocrine/paracrine mechanisms (Okumura et al., 1989). Thus, a mechanism different from that of EGFR transactivation can be envisioned here. The mitogenic activity of FGF is known to depend on its interaction with heparin and/or heparan sulfate (HS) proteoglycans that are strategically localized on the plasma membrane and in the extracellular matrix (Bashkin et al., 1989;Schlessinger et al., 2000; for review, see Ornitz 2000). FGF is secreted into the extracellular environment, where it can bind to heparin/HS, which coordinates its interaction with FGFR and prevents its diffusion and release into the interstitial space. Considerable evidence suggests that the extent and position of sulfation of HS dictates whether a stable ternary signaling complex is formed between a given FGF, heparin/HS, and its cognate FGFR or the secreted FGF leaves the cell surface (Schlessinger et al., 2000; for review, see Powers et al., 2000).
In the case of EGFR transactivation, it has been proposed that both PKC and Src may be direct regulators of the MMP that releases EGF-like ligand (Belcheva et al., 2001, and references therein). Although evidence for the involvement of an MMP for both MOR and KOR signaling to ERK was gained herein, a precise role for the protease in both pathways is unknown. A possible mechanism of this MMP potentiation for MOR is proteolytic, degradative remodeling of the extracellular matrix to expose HS proteoglycans to heparanase (Bashkin et al., 1989;Vlodavsky and Friedmann, 2001). A specific, hydroxamic acid-based MMP inhibitor (BB94) did not affect heparanase activity directly but potentiated heparanase-induced phenotypic changes to vascular smooth muscle cells (Fitzgerald et al., 1999). There are at least three mechanisms postulated to release FGF stored in the extracellular matrix so that it can bind to its receptor (Powers et al., 2000). One invokes an FGF-carrier binding protein to carry this growth factor to its receptor; another implicates heparanase in the release mechanism, whereas the third proposes that proteases release FGF-HS from the extracellular matrix (Vlodavsky and Friedmann, 2001). A question that remains then for FGFR1 transactivation concerns the identity of the proximal initiator and its target in the formation of a ternary complex between FGF, heparin/HS, and FGFR1. The MMP data suggest that both μ- and κ-opioid activation of ERK in C6 cells may involve the ectodomain shedding of growth factors. For MOR signaling, the data presented herein advocate the involvement of bFGF. Because KOR does not stimulate FGFR1 phosphorylation, it may induce the release of other growth factor(s) present in this cell line. Elucidation of the distinct differences in the mechanisms of μ- and κ-opioid signaling to ERK should shed light on their physiological and pathophysiological actions along with those of bFGF.
Acknowledgments
We thank Dr. David Ornitz (Department of Pharmacology and Molecular Biology, Washington University, St. Louis, MO) for the dominant negative mutant of FGFR1.
Footnotes
- Received May 8, 2002.
- Accepted August 9, 2002.
This study was supported by National Institutes of Health Grant DA05412.
DOI: 10.1124/jpet.102.038554
Abbreviations
- GPCR
- G protein-coupled receptor
- RTK
- receptor tyrosine kinase
- ERK
- extracellular signal-regulated kinase
- MOR
- μ-opioid receptor
- LPA
- lysophosphatidic acid
- EGFR
- epidermal growth factor receptor
- HEK
- human embryonic kidney
- MMP
- matrix metalloproteinase
- EGF
- epidermal growth factor
- PDGRF
- platelet-derived growth factor receptor
- FGF
- fibroblast growth factor
- IGF-1
- insulin-like growth factor-1
- PDGF
- platelet-derived growth factor
- bFGF
- basic fibroblast growth factor
- KOR
- κ-opioid receptor
- PTX
- pertussis toxin
- PKC
- protein kinase C
- FGFR
- fibroblast growth factor receptor
- DAMGO
- [d-Ala2,N-Me-Phe4,Gly5-ol]-enkephalin
- CTAP
- d-Phe-Cys-Tyr-d-Trp-Arg-Thr-Pen-Thr-NH2
- DMEM
- Dulbecco's modified Eagle's medium
- MEM
- minimal essential medium
- CS
- calf serum
- PMA
- phorbol 12-myristate 13-acetate
- BAPTA
- 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid
- GFX
- bisindolylmaleimide I
- βARK
- β-adrenergic receptor kinase
- PI3K
- phosphatidylinositol 3-kinase
- VEGFR
- vascular endothelial growth factor receptor
- HS
- heparan sulfate
- U69,593
- (5α,7α,8β)-(−)-N-methyl-(7-(1-pyrrolidinyl)-1-oxospiro(4,5)dec-8-yl)-benzeneacetamide
- SU5416
- 3-[(2,4-dimethylpyrrol-5-yl)methylidenyl]indolin-2-one
- SU6668
- (Z)-3-[2,9-dimethyl-5-(2-oxo-1,2-dihydro-indol-3-ylidenemethyl)-1H-pyrrol-3-yl]propionic acid
- The American Society for Pharmacology and Experimental Therapeutics