Morphine-induced signaling via opioid receptors (ORs) in dorsal root ganglia (DRG) neurons, the spinal cord, and various brain regions has been shown to modulate gene activity. Hitherto, little attention has been paid to extracellular signal-regulated kinases-1/2 (ERK-1/2)-mediated activation of the serum response factor (SRF) and ternary complex factors (TCFs) such as the E twenty six-like transcription factor-1 (ELK-1) in this context. Using TCF/SRF-dependent reporter gene constructs, a specific ERK-1/2 inhibitor and a dominant-negative ELK-1 mutant, we show herein that morphine activates ELK-1 via ERK-1/2 in DRG-derived F11 cells endogenously expressing μ and δ ORs. Previous studies with glioma cell lines such as NG108-15 cells attributed morphine-induced gene expression to the activation of the cAMP-responsive element binding protein (CREB). Thus, we also analyzed morphine-dependent activation of CREB in F11 and NG108-15 cells. In contrast to the CREB stimulation found in NG108-15 cells, we observed an inhibitory effect of morphine in F11 cells, indicating cell type-specific regulation of CREB by morphine. To obtain data about putative target genes of morphine-induced ELK-1/SRF activation, we analyzed mRNA levels of 15 ELK-1/SRF-dependent genes in cultured rat DRG neurons and F11 cells. We identified the early growth response protein-4 (EGR-4) as the strongest up-regulated gene in both cell types and observed ELK-1 activity-dependent activation of an EGR-4-driven reporter in F11 cells. Overall, we reveal an important role of ELK-1 for morphine-dependent gene induction in DRG-derived cells and propose that ELK-1 and EGR-4 contribute to the effects of morphine on neuronal plasticity.
Opioid receptors (ORs) belong to the superfamily of G protein-coupled receptors (GPCRs) and mediate their biological effects via the activation of pertussis toxin (PTX)-sensitive Gi/o proteins (Parolaro et al., 1990). Application of the μ OR (MOP)-selective agonist morphine significantly reduces nociception but also alters gene activity, leading to dependence and tolerance (Li and Clark, 1999; Ammon-Treiber and Höllt, 2005). Spinal cord or brain neurons have been intensively analyzed in this regard (Taylor and Fleming, 2001), but despite the important role of primary sensory or dorsal root ganglia (DRG) neurons in the analgesic actions of morphine, the effects of morphine on gene expression in these neurons are less understood.
GPCR-promoted signaling increases gene activity via cAMP response elements (CRE) or serum response elements (SRE). CRE is activated because of its interaction with the CRE binding protein (CREB) (Andrisani, 1999), whereas SRE activity is enhanced after binding to the serum response factor (SRF) and ternary complex factors (TCFs) such as E twenty six (ETS)-like transcription factor-1 (ELK-1) (Buchwalter et al., 2004). Interactions between CRE and CREB are enhanced after the phosphorylation of CREB by numerous downstream kinases of GPCR signaling such as protein kinase A (PKA), protein kinase C (PKC), calmodulin-dependent kinase II (CamKII), and extracellular signal-regulated kinases-1/2 (ERK-1/2) (Shaywitz and Greenberg, 1999). Likewise, affinity of the ELK-1/SRF/SRE complex is increased after phosphorylation of ELK-1 by ERK-1/2 (Davis, 1995).
The effects of morphine on gene regulation have been attributed to the phosphorylation of CREB and the concomitant activation of CRE (Maldonado et al., 1996; Widnell et al., 1996; Li and Clark, 1999; Zhou and Zhu, 2006). Glioma-derived NG108-15 or Neuro2A cells have extensively been used to study opioid-induced signaling (Blume, 1978; Chakrabarti et al., 1995). In these cell lines morphine exerts its effects on CRE via CamKII- or PKC-mediated phosphorylation of CREB (Bilecki et al., 2000, 2004). However, given that opioids engage signaling pathways that both increase (activation of ERK-1/2, CamKII, and/or PKC) and decrease CREB activity (PKA inhibition via Gi/o), the net effect of such opposing morphine effects on CRE activity might differ in distinct cell types. ERK-1/2 play a central role in GPCR-dependent gene induction, and downstream targets of ERK-1/2 such as ELK-1 have been reported to play an important role in the gene regulation of adult DRGs (Kerr et al., 2010) and mediate the central effects of other abusive drugs such as cocaine (Besnard et al., 2011). Hence the ERK-1/2/ELK-1/SRF pathway emerges as a likely alternative by which morphine could induce gene expression in the absence of CREB activation.
So far, surprisingly little is known about the role of the ERK-1/2/ELK-1/SRF pathway in morphine-induced gene expression. In MOP-overexpressing cells inconsistent observations have been made, because both etorphine and morphine induced ERK-1/2 phosphorylation in HEK293 cells, but only etorphine activated an ELK-1-driven reporter gene construct (Zheng et al., 2008), whereas in MOP-overexpressing Chinese hamster ovary cells morphine-induced activation of an ELK-1 reporter was observed (Shoda et al., 2001). Signaling pathways responsible for ELK-1 activation in different OR-overexpressing cell models have not been sorted out, and to our knowledge activation of the ELK-1/SRF pathway by morphine in DRG neurons or other endogenous expression systems has not been described.
F11 cells are regarded as an authentic cell model for DRG neurons (Francel et al., 1987) and endogenously express δ OR (DOP) and MOP subtypes (Fan et al., 1993). Thus, to investigate morphine-dependent gene induction in a DRG-derived cell line and answer the question of whether morphine regulates ELK-1/SRF in an endogenous cell system, we analyzed the effects of opioids on various reporter gene constructs in F11 cells. First, we found that morphine significantly enhances ELK-1/SRF signaling via ERK-1/2 in F11 cells, and second, comparing the CREB/CRE pathway in NG108-15 and F11 cells, we observed that morphine increases CREB activity in the former, but depresses its activity in the latter cell model. To get an idea about the putative physiological role of morphine-induced ELK-1/SRF activation, we also analyzed mRNA levels of 15 ELK-1/SRF-dependent genes in cultured rat DRG neurons and F11 cells. We identified the early growth response protein-4 (EGR-4) as the gene most strongly up-regulated in both cell types. In line with morphine-promoted EGR-4 mRNA induction via ERK-1/2, we observed ELK-1 activity-dependent activation of an EGR-4-driven reporter by morphine in F11 cells. Overall, we propose that the ELK-1/SRF pathway plays a predominant role in morphine-promoted gene induction in DRG-derived cells and define EGR-4 as a novel target of morphine-promoted signaling.
Materials and Methods
HAM's F12 medium, DMEM, fetal bovine serum, penicillin/streptomycin, and HAT were purchased from Invitrogen (Darmstadt, Germany). [3H]naloxone was from PerkinElmer Life and Analytical Sciences (Waltham, MA). The transfection reagent PromoFectin was from PromoKine (Heidelberg, Germany). The primary antibodies p-ERK-1/2 and t-ERK2 were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), and p-ELK-1 was from Cell Signaling Technology (Danvers, MA). Anti-mouse or anti-rabbit horseradish peroxidase-conjugated secondary antibodies were purchased from Bio-Rad Laboratories (Munich, Germany). PTX and morphine were obtained from Sigma-Aldrich (Deisenhofen, Germany). DAMGO, DPDPE, CTAP, and 2-(2-chloro-4-iodophenylamino)-N-cyclopropylmethoxy-3,4-difluorobenzamide (PD-184352) were purchased from Tocris Bioscience (Bristol UK). Morphine sulfate dissolved in methanol was from Sigma-Aldrich (St. Louis, MO).
Cell Culture and Transfection.
F11 cells were cultured in HAM's F12 medium (15% fetal bovine serum, 2 mM l-glutamine, HAT, 100 U/ml penicillin, and 100 μg/ml streptomycin). NG108-15 cells (kindly provided by Dr. Daniela Eisinger, Ludwig-Maximilians University, Munich, Germany) were cultured in DMEM (10% fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, and HAT). For transfection, cells were seeded at a density of 2 × 106 cells in a 10-cm dish and transfected with the PromoFectin reagent according to the manufacturer's protocol. For reporter gene assays or detection of ERK-1/2 or ELK-1 phosphorylation, cells were serum-starved for 16 h. Cryo-preserved DRG neurons from day-18 embryonic Sprague/Dawley rats were purchased from Innoprot (Derio, Spain). After isolation from the animal, DRG neurons were cultured for 5 days at Innoprot, detached from the cell culture plate, cryo-preserved, shipped on dry ice, and then cultured in our laboratory for an additional 7 to 10 days on poly-l-ornithin-coated six-well plates in neurobasal medium containing 5 mM glutamine, 10% serum, B27 supplement, and penicillin/streptomycin/neomycin. Media were replaced every other day. One day before stimulation DRG neurons were starved by replacing complete medium with neurobasal medium containing only glutamine.
Radioligand Binding Assay.
For ligand binding studies, cells were detached and washed with HAM's F12 medium containing 100 mM HEPES, pH 7.4, and 0.1% BSA. Approximately 25,000 cells were then incubated with various concentrations of [3H]naloxone in the absence or presence of unlabeled naloxone (10 μM) in ice-cold HAM's F12 containing 100 mM HEPES, pH 7.4, and 0.1% BSA. Radioligand binding assays were carried out for 1 h, and the amount of cell-bound radioactivity was determined in a β-counter after washing and collecting cells in a cell harvester.
Aequorin-Based Calcium Measurements.
Twenty four hours after transfection of ∼2 × 106 F11 cells with a calcium-sensing aequorin-eGFP construct (pG5α; kindly provided by Dr. Vladimir Chubanov, Ludwig-Maximilians University) in a 10-cm dish, cells were loaded with the aequorin substrate coelenterazine H (5 μM) in HBS buffer for 30 min at room temperature. After cell harvesting in HBS without BSA, ∼1 × 105 cells per well were seeded in 96-well plates, and total luminescence was measured in a FLUOstar Omega plate reader (BMG, Offenburg, Germany) at 37°C. HBS, 0.02% methanol HBS, or HBS with the corresponding ligand was automatically injected after 5 s. Total emission was measured at 1-s intervals. Luminescence was normalized by defining the first ratio (1 s) measured as 100%.
To determine agonist-promoted cAMP accumulation ∼200,000 cells were seeded in 12-well plates coated with 0.1% poly-l-lysine 24 h before the experiment and labeled in serum-free HAM's F12 medium containing 2 μCi/ml of [3H]adenine for 16 h. Cells were stimulated for 30 min at 37°C in DMEM containing 1 μM 3-isobutyl-1-methylxanthin, forskolin (FSK) when indicated, and various concentrations of different ligands. The reaction was terminated by removing the medium and adding ice-cold 5% trichloroacetic acid. [3H]ATP and [3H]cAMP were then purified by sequential chromatography using dowex-resin/aluminum oxide columns, and radioactivity was measured in a β-counter.
Firefly Luciferase Reporter Gene Assays.
For the detection of TCF/SRF activity a reporter gene construct containing an AP-1 site (TGAGTCA) for c-Fos/c-jun, a cis-inducible element site (TTCCCGTCAA) for STAT and an ETS/SRE site (GGATGTCCATATTAGGACATC) for TCF/SRF, or a reporter solely containing five repeats of the ETS/SRE-site were used. Note that the ETS site binds ELK-1, SRF accessory protein-1, and neuroepithelial transforming gene-1. Twenty four hours after transfection, cells were seeded in 24-well plates coated with 0.1% poly-l-lysine. After 24 h, cells were serum-starved for 16 h and then stimulated in serum-free HAM's F12 medium or DMEM containing various ligands for 6 h. The dominant negative ELK-1 mutant (REST/ELK-1-δC) has been described previously (Stefano et al., 2007) and was kindly provided by Dr. Gerald Thiel (Universität des Saarlandes, Homburg, Germany). To monitor CRE activity, a pAD-CRE-firefly luciferase (Fluc) plasmid containing six CRE sites was transfected into F11 or NG108-15 cells. The p0.2NxL reporter (kindly provided by Dr. Refugio García-Villegas, Cinvestav-Pin, Mexico City, Mexico) has been described previously (García-Villegas et al., 2009) and contains the coding sequence of the Fluc under the control of the minimal promoter of the mouse sodium-activated sodium channel type VII α (Scn7a) that harbors a binding site for EGR-4 and an E box. Cells were lysed, and Fluc activity was determined by using a luciferase reporter system (Promega, Mannheim, Germany) according to the manufacturer's protocol by using a FLUOstar Omega plate reader (BMG). If reporter gene activity of distinct transfection was compared, overall transfection efficacies were controlled by con-expression of a plasmid encoding soluble YFP under the control of a constitutively active promotor. YFP emission (excitation, 480 ± 10; emission, 520 ± 10) was determined in the lysates by using the FLUOstar Omega.
Western Blot Analysis.
Forty eight hours before the experiment, ∼4 × 105 F11 cells were seeded in six-well plates. After 1 day cells were serum-starved for 16 h. Stimulation of cells was carried out for the indicated time periods at 37°C and stopped by cooling on ice. Cells were washed once with ice-cold phosphate-buffered saline and lysed in Laemmli buffer. Lysates were homogenized by sonication, heated for 5 min at 95°C, and centrifuged (17,000g; 5 min). Proteins were resolved by SDS-polyacrylamide gel electrophoresis and subjected to Western blotting by using various antibodies, p-ERK1/2 (1:2000) from Santa Cruz Biotechnology (E-4), t-ERK2 antibody (1:10,000) from Santa Cruz Biotechnology (C-14), or p-ELK1 (Ser383) (1:5000) from Cell Signaling Technology and the corresponding horseradish peroxidase-coupled secondary antibody (1:10,000) from Bio-Rad Laboratories, goat anti-mouse IgG, and goat anti-rabbit IgG. Immune reactivity was detected by monitoring enhanced chemiluminescence-dependent light emission with a chemiluminescence detection system (Peqlab, Erlangen, Germany). Nitrocellulose membranes were stripped with 0.2 M glycine, pH 2 at 37°C for 48 h and then reprobed.
Serum-starved F11 cells or cultured rat DRG neurons were stimulated for 1 h with 1 μM morphine in serum-free medium. PD-184352 (10 μM) was preincubated for 30 min. Stimulation was terminated by rapid cooling on ice, and total RNA was isolated by using the TriFast reagent (Invitrogen) according to the manufacturer's instructions. First-strand synthesis was carried out with oligo(dT)18 primer using 2 μg of total RNA and the RevertAid H Minus First Strand cDNA Synthesis Kit (Fermentas, Sankt-Leon Roth, Germany). qRT-PCR was done by using the LightCycler 480 SybrGreen I Master Mix (Roche Diagnostics, Mannheim, Germany), intron-spanning primer pairs at a final concentration of 1 μM each, and 0,08 μl of the first-strand synthesis reaction in a LightCycler 480 (Roche Diagnostics) by using the following conditions: initial denaturation for 15 min at 94°C, 55 cycles of 94°C for 10 s, 55°C for 10 s, and 72°C for 10 s. Crossing points (Cp) were determined by the software supplied with the LightCycler 480. Sequences of primer pairs used are given in Table 1.
Data obtained by Western blotting were analyzed by using ImageJ (National Institutes of Health, Bethesda, MD). Statistical significance of differences was assessed by Student's t test (between two groups) or one-way analysis of variance and Tukey's honest significance post hoc test (between more than two groups).
Opioid Receptor Expression in F11 Cells.
F11 (rat dorsal root ganglia neurons × mouse neuroblastoma) cells are generally regarded as a suitable cell model for analyzing DRG neuron-specific signaling (Francel et al., 1987; Boland et al., 1991; Puttfarcken et al., 1997; McIlvain et al., 2006; Jung and Miller, 2008) and endogenously express DOP and MOP but not κ opioid receptor subtypes (Fan et al., 1992, 1993). MOP-specific agonists such as DAMGO increase intracellular calcium concentrations in these cells. In line with these prior findings, we detected 367 ± 92 fmol/mg specific OR binding sites in the F11 cell pool studied (Fig. 1A) and observed that DAMGO (Fig. 1B) or morphine (Fig. 1C) induced intracellular calcium transients. Thus, F11 cells may serve as a suitable model system for analyzing the effects of morphine on the ELK-1/SRF pathway.
Morphine-Induced TCF/SRF-Driven Protein Expression in F11 Cells.
ERK-1/2 have been shown to be targets of morphine-dependent signaling and play a critical role in gene induction, because they are dominant upstream regulators of TCFs and CREB. Thus, we first analyzed effects of morphine on ERK-1/2 phosphorylation. We observed a rapid and transient increase in ERK-1/2 phosphorylation with a peak at 2.5 min and a subsequent decline thereafter (Fig. 2, A and C). In accord with previous studies (Parolaro et al., 1990), Gi/o proteins are required for morphine-induced signaling, because treatment of F11 cells with PTX completely blocked morphine-promoted phosphorylation of ERK-1/2 (Fig. 2, B and C).
To investigate morphine-dependent gene induction in F11 cells, we took advantage of a reporter gene construct that encodes a promoter containing cis elements for STAT, TCF/SRF, and AP-1 (for details see Materials and Methods). Expression of this reporter construct revealed that DAMGO, morphine, and the DOP-specific agonist DPDPE induced luciferase expression to an extent comparable with serum stimulation (Fig. 3A), indicating that morphine enhanced gene expression in F11 cells via STAT, TCF/SRF, or AP-1. To narrow down the cis element responsible for morphine-induced reporter activation, we next expressed a reporter gene construct containing five repeats of TCF/SRF binding sites, but no consensus sites for STAT or AP-1. Morphine induced significant luciferase expression via TCF/SRF similar to the DAMGO effect and more pronounced than the gene expression induced by DPDPE (Fig. 3B). Thus, we provide the first evidence for morphine-induced gene expression via TCF/SRF in neuronal, DRG-derived cells. Morphine activates both MOP and DOP, but exhibits selectivity toward the MOP subtype (Fukuda et al., 1995). Accordingly, the morphine concentration used (1 μM) was assumed to activate mainly the MOP. To test this hypothesis, we applied CTAP, a highly selective MOP antagonist (Pelton et al., 1986), and measured ligand-induced TCF/SRF reporter activity. Indeed, we observed a significant reduction in DAMGO- and morphine-, but not DPDPE-induced reporter activity (Fig. 3C), indicating that morphine mediates its effects in F11 cells via the MOP subtype.
Morphine-Induced Activation of TCF/SRF in F11 Cells Required ERK-1/2 and ELK-1 Activation.
The ETS site of the TCF/SRF-driven reporter used herein contains a binding site for ELK-1, SRF accessory protein-1, or neuroepithelial transforming gene-1 (Price et al., 1995). In addition to ERK-1/2, other kinases such as c-Jun N-terminal kinase have been shown to activate TCFs (Yordy and Muise-Helmericks, 2000), and even TCF-independent SRF activation mediated by the monomeric G protein Rho and myocardin-related transcription factors has been demonstrated (Hill et al., 1995; Posern and Treisman, 2006). Therefore, to further dissect the signaling pathway underlying morphine-induced activation of the TCF/SRF reporter and elucidate the role of ERK-1/2, we tested the impact of an ERK-1/2 inhibitor (PD-184352) and a dominant-negative ELK-1 mutant (REST/ELK-1-δC) on morphine-dependent TCF/SRF reporter activation (Stefano et al., 2007). PD-184352 significantly inhibited basal, serum and even stronger morphine-promoted reporter activity (Fig. 4A), indicating that ERK-1/2 have a major contribution to the effects of morphine on the TCF/SRF reporter. In line with this notion, we observed increased phosphorylation of ELK-1 after 2 to 5 min (first peak) and 20 to 30 min (second peak) of morphine stimulation, an effect that was again blocked by PD-184352 (Fig. 4B). Coexpression of the REST/ELK-1-δC mutant and the TCF/SRF reporter significantly inhibited basal reporter activity, although coexpression of the ELK-1 mutant did not affect the overall transfection efficiency of a plasmid encoding soluble YFP (Fig. 4C), indicating that ELK-1 activity is generally required for TCF/SRF-driven reporter activity in F11 cells. In addition, depressing ELK-1 activity by expression of a dominant negative mutant precluded morphine-promoted ETS/SRE activation (Fig. 4C), indicating that ELK-1 is required for morphine-promoted activation of the TCF/SRF reporter.
Cell Type-Specific CREB Regulation by Morphine in F11 and NG108-15 Cells.
Having established that morphine initiates gene induction via ERK-1/2 and the ELK-1/SRF/SRE pathway, we wondered whether morphine would additionally activate CREB and CRE in F11 cells. Because cAMP-dependent protein kinase A (PKA) is the major activator of CREB, we measured the effects of morphine on forskolin-induced cAMP accumulation and found that DAMGO, morphine, and DPDPE significantly reduced FSK-induced cAMP accumulation in a PTX-sensitive manner (Fig. 5A). To assess the influence of morphine-dependent inhibition of cAMP production on CREB activity, we next used a reporter gene construct that contained six CRE sites. Although we did not observe increased CREB activity in response to a 4-h stimulation by DAMGO or morphine, the CRE reporter strongly reacted to FSK, which increases cAMP levels independently of OR (Fig. 5B). In contrast, 16 h after ligand stimulation significant inhibition of CREB by morphine and DAMGO was detectable (Fig. 5D). These data contrast with previously observed opioid-induced phosphorylation/activation of CREB in other cell lines (Bilecki et al., 2000, 2004). Therefore, we included NG108-15 cells in our experiments. We noted that in analogy to the situation in F11 cells DAMGO, morphine, and DPDPE significantly reduced cAMP levels in NG108-15 cells (Fig. 5C). However, in accord with previous reports, expression of the DOP subtype is predominant compared with the MOP (Gomes et al., 1999). It is noteworthy that in NG108-15 cells morphine transiently increased CRE activity after 4 h, but did not show a stimulatory effect after 16 h (Fig. 5, B and D), suggesting cell type-specific differences in the regulation of CREB. Next, we analyzed the effects of PD-184352 and BIM-X (a PKC inhibitor) on morphine-promoted CRE activity. Whereas PD-184352 further enhanced the inhibitory effects of morphine on CRE in F11 cells (Fig. 5E), BIM-X blunted the activating effects in NG108-15 cells (Fig. 5F), indicating that ERK-1/2 may contribute to CRE activation by morphine in the former cell model, whereas PKC is indispensable for CRE stimulation in the latter. However, ERK-1/2-mediated CRE activation seems to be masked by dominant Gi/o-mediated inhibition of the cAMP/PKA/CREB pathway by morphine in F11 cells (summarized in Fig. 9).
Morphine Does Not Activate the ELK-1/SRF Pathway in NG108-15 Cells.
Given cell type-specific differences in morphine-promoted activation of CREB, we next wondered whether similar to the situation in F11 cells morphine activates the TCF/SRF pathway in NG108-15 cells. As shown in Fig. 6A, stimulation of NG108-15 cells with 1 μM or even 10 μM morphine failed to induce TCF/SRF reporter activity, whereas serum and DPDPE significantly activated the reporter, indicating cell type-specific differences in morphine-promoted TCF/SRF activation. This is of particular interest because 10 μM morphine similarly affected cAMP production compared with 1 μM DPDPE (Fig. 6B), indicating that the inability of morphine to activate TCF/SRF in NG108-15 cells is not simply caused by its weaker potency to activate the DOP.
Morphine Induces Expression of Early Growth Response Protein-4 in Cultured Rat DRG Neurons and F11 Cells.
So far, our data suggest that morphine-dependent gene induction is mediated primarily by ERK-1/2 and the ELK-1/SRF/SRE complex in F11 cells. To gain first insight into a putative physiological role of this pathway, we aimed to identify target genes that are up-regulated via ELK-1/SRF in DRG neurons after morphine stimulation. In detail, we determined mRNA levels of 15 genes known to be regulated by SRF in neurons (Knöll and Nordheim, 2009) by performing qRT-PCR experiments with cDNAs obtained from cultured rat DRG neurons. As summarized in Table 1 and Fig. 7A, stimulation of DRG neurons with morphine (1 μM) for 1 h significantly up-regulated mRNA levels of Fos-B, SRF, and EGR-4, with EGR-4 being the strongest induced gene. To verify the role of ERK-1/2 in this process, we analyzed the effects of PD-184352 on morphine-promoted EGR-4 induction. The ERK-1/2 inhibitor completely blocked EGR-4 induction (Fig. 7C), indicating an important role for ERK-1/2 in this process. Next, we performed identical qRT-PCR experiments with F11 cells and found that similar to DRG neurons morphine significantly induced mRNA expression of Fos-B and EGR-4 and additionally of EGR-2 and CRE modulator-2 (CREM-2) (Table 1; Fig. 7B). It is noteworthy that similar to DRG neurons EGR-4 appeared as the strongest induced gene, indicating similar effects of morphine on gene expression in F11 cells and cultured DRG neurons. In addition, compatible with our data obtained with DRGs, PD-184352 also blocked EGR-4 induction in F11 cells (Fig. 7D). Thus ERK-1/2-mediated expression of EGR-4 by morphine seems to be a common event in DRGs and DRG-derived cells.
Morphine-Promoted Activation of an EGR-4-Driven Reporter Requires ELK-1 Activity in F11 Cells.
To validate our data obtained with qRT-PCR experiments, we next set out to analyze the effects of morphine on EGR-4-promoted gene induction. It has been reported that expression of the sodium-activated sodium channel Scn7a in DRG neurons is regulated by a minimal promoter containing a binding site for EGR-4 (García-Villegas et al., 2009). Thus, we took advantage of a reporter gene construct (p0.2NxL) that contains this promoter sequence and analyzed the effects of morphine on EGR-4-driven reporter activity in F11 cells. We found that morphine transiently increased EGR-4-dependent reporter activity (Fig. 8), indicating that morphine induced the expression of biologically active EGR-4. The promoter of the EGR-4 gene has been reported to contain CRE and SRE sites (Crosby et al., 1992; Holst et al., 1993). To narrow down the cis element responsible for morphine-promoted EGR-4 expression in F11 cells, we coexpressed the dominant negative mutant of ELK-1 and the p0.2NxL reporter. Morphine failed to increase EGR-4-dependent protein expression under these conditions (Fig. 8), indicating that ELK-1 activation is responsible for morphine-promoted EGR-4 induction in F11 cells. Thus, we identified EGR-4 as the first gene to be up-regulated by morphine via ELK-1 in DRG-derived cells.
Morphine has been shown to induce cellular adaptations that occur on the posttranslational or transcriptional level and lead to tolerance and dependence (Li and Clark, 1999; Gintzler and Chakrabarti, 2000; Chakrabarti et al., 2001; Ammon-Treiber and Höllt, 2005). Activation of CREB and ELK-1/SRF are probably the most common pathways by which GPCR-activating drugs affect gene expression. Although ELK-1 and SRF have been shown to modulate neuronal plasticity (Yu and Yezierski, 2005; Ji et al., 2009; Knöll and Nordheim, 2009; Besnard et al., 2011; Klinger et al., 2011), surprisingly little attention has been paid to ELK-1/SRF in the context of morphine-induced neuronal plasticity, which contributes to drug tolerance/dependence. Therefore, we analyzed the effects of morphine on ELK-1/SRF-mediated gene regulation in F11 cells endogenously expressing the DOP and MOP subtype. We observed morphine-induced activation of the ELK-1/SRF pathway and thus show for the first time that morphine is able to modulate gene expression via ELK-1/SRF in an endogenous cell system. Because morphine failed to activate the CREB/CRE pathway, our findings highlight a unique role of ELK-1 for morphine-induced gene expression in F11 cells and separate these cells from established cell models such as NG108-15 cells that show morphine-promoted CREB activation and no effects on ELK-1/SRF (Fig. 9).
Considering putative signaling pathways responsible for morphine-induced activation of ELK-1/SRF in F11 cells, it is noteworthy that in MOP-overexpressing HEK293 cells opioids have been shown to activate ERK-1/2 via PKC and β-arrestins (Zheng et al., 2008). PKC-dependent activation of ERK-1/2 occurred rapidly (2–5 min) and did not alter ELK-1 activity, whereas β-arrestin-promoted activation of ERK-1/2 occurred at later time points (20–30 min) and increased ELK-1-driven reporter activity. Commensurate with the established inability of morphine to recruit β-arrestins to the MOP in HEK293 cells or other cell lines (Keith et al., 1996), it consequently failed to activate ELK-1 despite its robust PKC-mediated activation of ERK-1/2. In F11 cells, inhibition of PKC activity by BIM-X failed to block morphine-promoted ERK-1/2 activation (data not shown), suggesting that ERK-1/2 and thus also ELK-1/SRF are differentially regulated in HEK293 and F11 cells. Given that morphine-promoted ERK-1/2 activation occurred rapidly (after 2.5 min), we assume that β-arrestins are not involved. However, further studies are required to determine a putative role for β-arrestins in morphine-promoted ELK-1 activation in F11 cells. At this point, we postulate that unidentified, PTX-sensitive signaling pathways are responsible for the morphine-promoted ELK-1/SRF activation via ERK-1/2 in F11 cells.
Cell type-specific gene regulation by morphine raises an important question about the molecular determinants responsible for these differences. It is noteworthy that F11 cells express DOP and MOP but not κ opioid receptor subtypes (Fan et al., 1993), whereas NG108-15 cells almost exclusively express the DOP (Gomes et al., 1999). In line with this overall situation, the effects of DAMGO on cAMP production were much more pronounced in F11 compared with NG108-15 cells. Thus, cell type-specific activation of ELK-1/SRF might reflect exclusive morphine-induced ELK-1/SRF activation by MOP receptors. This assumption is supported by the finding that a MOP-specific antagonist blocked morphine-induced ELK-1/SRF activation in F11 cells. However, the DOP-specific agonist DPDPE was able to activate ELK-1/SRF in NG108-15 cells, indicating that DOPs are functionally linked to this pathway. Because 10 μM morphine did not activate ELK-1/SRF, although they similarly affected cAMP production compared with 1 μM DPDPE, it seems that DOP-mediated activation of ELK-1/SRE is distinctly regulated by different agonists in NG108-15 cells. Future investigations are required to clarify the exact mechanisms responsible for ligand-specific activation of ELK-1/SRF via DOP in NG108-15 cells.
With regard to cell type-specific morphine-promoted signaling, it is also of note that fusion of neuroblastoma NG18TG-2 cells with rat DRG neurons created the F11 cell line (Fan et al., 1992) and with glioma cells (C6-BU-1 cells) created the NG108-15 cell line (Augusti-Tocco and Sato, 1969). Regulation of gene activity might generally differ in cells of different origins. Fundamental differences among both cell types are first reflected by our finding that a PKC inhibitor reduced morphine-induced CREB activation in NG108-15 but not in F11 cells, whereas inhibition of ERK-1/2 affected morphine-promoted CREB activation in the latter but not in the former cell line. Second, it was remarkable that short FSK stimulation (4 h) activated CRE stronger in F11 cells, and prolonged stimulation (16 h) was more efficacious in NG108-15 cells (Fig. 5, B and D), suggesting that PKA regulates CREB activity in NG108-15 cells with slower kinetics compared with F11 cells. In line with its selectivity toward the MOP subtype, the impact of morphine (1 μM) on the cAMP pathway was more robust in F11 than in NG108-15 cells. Thus, the weak effects of morphine on cAMP/PKA in NG108-15 cells might allow morphine-promoted PKC activation to activate CREB, despite the inhibitory actions of morphine on CREB caused by PKA inhibition. In contrast, the fast and strong effects of morphine on cAMP/PKA sufficiently inhibited CREB activity and thus obscured morphine-promoted CREB activation via ERK-1/2 in F11 cells. Therefore, we postulate that weak and belated effects of morphine on PKA activity are responsible for transient CRE activation in NG108-15 cells, whereas profound and fast effects of morphine on PKA curtail CRE activation in F11 cells (Fig. 9).
Overall, we provide evidence that morphine increases ELK-1/SRF-dependent gene expression in an endogenous cell system derived from DRG neurons. In line with this notion, morphine increased mRNA levels of Fos-B and EGR-4 (alias NGFI-C or pAT133) not only in F11 cells but also in cultured embryonic DRG neurons. The Fos-B and the EGR-4 promoter has been reported to contain SRE and CRE sites (Crosby et al., 1992; Lazo et al., 1992; Holst et al., 1993). Thus, morphine-induced expression of both proteins could be mediated either by TCF/SRF or CREB. Because morphine did not activate CRE and a dominant-negative ELK-1 mutant blocked morphine-induced activation of an EGR-4-driven reporter in F11 cells, we propose that ERK-1/2-mediated activation of the ELK-1/SRF/SRE pathway is responsible for the effects of morphine on EGR-4 expression.
With regard to a putative physiological relevance of morphine-induced effects on ELK-1 and CREB, it is interesting that previous studies revealed a switch from cell proliferation to differentiation associated with neurite outgrowth in F11 cells after ELK-1 inhibition and concomitant CREB activation (Ghil et al., 2000; McIlvain et al., 2006). Therefore, morphine has the propensity to counteract neurite outgrowth in F11 cells by activating ELK-1/SRF and inhibiting CREB. This is of further interest when considering morphine-induced expression of the CREM-2 in F11 cells (Fig. 7B), which has been reported to inhibit CREB (Foulkes et al., 1991), and thus could further enhance the inhibitory effects of morphine on cell differentiation.
Considering the putative physiological relevance of our findings in terms of morphine tolerance/dependence, it is noteworthy that Fos-B and its stable isoform ΔFos-B have been recognized as sustained molecular switches of addiction to cocaine and other abusive drugs in brain neurons (Nestler et al., 1999, 2001), suggesting that the induction of Fos-B by morphine in DRG neurons could have similar effects. However, at the level of DRG neurons it is more likely that proteins that modulate the excitability of primary sensory neurons and/or signal transmission from these neurons to the spinal cord are involved in morphine tolerance. EGR-4 is an autoinhibitory transcription factor with a zinc finger domain (Zipfel et al., 1997) that has been reported to induce the expression of the sodium channel Scn7a in DRG neurons (García-Villegas et al., 2009) and the neuron-specific K+/Cl− cotransporter-2 (KCC-2) in the hippocampus and cerebellum (Uvarov et al., 2006). KCC-2 transporters enhance synaptic transmission in particular by inhibitory transmitters such as glycine and γ-aminobutyric acid (Rivera et al., 1999; Hübner et al., 2001), suggesting that putative morphine-promoted KCC-2 induction would alter neuronal transmission via inhibitory transmitters between DRGs and the spinal cord. Increased sodium influxes into DRGs caused by enhanced Scn7a expression would lower the resting potential of these neurons, thus increasing their excitability. In such a scenario, inhibitory effects of morphine on neuronal excitability caused by the activation of potassium channels and inhibition of voltage-gated calcium channels would be diminished, so that increased morphine concentrations would be necessary to obtain similar analgetic actions, leading to drug tolerance.
In conclusion, we reveal an as-yet-unappreciated role of the ELK-1/SRF pathway in morphine-promoted gene induction. It seems that this new pathway plays a dominant role in DRG-derived cells and links MOP subtypes to other transcription factors such as Fos-B and EGR-4. EGR-4 in turn induces the expression of neuron-specific proteins such as sodium channels or ion transporters and thus could contribute to the development of drug tolerance or dependence. Further studies are required to sort out whether the up-regulation of ELK-1/SRF-dependent genes by morphine is restricted to DRG-derived cells and how this pathway affects the sensitivity of primary sensory neurons to morphine or other stimuli in vivo.
Participated in research design: Rothe, Boekhoff, Gudermann, and Breit.
Conducted experiments: Rothe, Solinski, and Breit.
Contributed new reagents or analytic tools: Breit.
Performed data analysis: Rothe, Boekhoff, and Breit.
Wrote or contributed to the writing of the manuscript: Gudermann and Breit.
We thank Dr. Susanne Mühlich (Walther-Straub-Institute for Pharmacology and Toxicology, Munich, Germany) for the SRE reporter; Dr. Vladimir Chubanov (Walther-Straub-Institute for Pharmacology and Toxicology) for the pG5α aequorin-eGFP construct; Dr. Gerald Thiel (Universität des Saarlandes, Homburg, Germany) for the REST/ELK-1-δC mutant; Dr. Refugio García-Villegas (Cinvestav-Pin, Mexico City, Mexico) for the p0.2NxL reporter; and Ute Künzel-Mulas for excellent technical assistance.
This work was supported by the “FöFoLe” program of the Medicine Department of Ludwig-Maximilians University of Munich [Grant 43/2009].
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- opioid receptor
- δ OR
- μ OR
- activating protein-1
- bisindolylmaleimide X
- bovine serum albumin
- calmodulin-dependent kinase II
- crossing point
- cAMP response element
- CRE binding protein
- CRE modulator
- Dulbecco's modified Eagle's medium
- dorsal root ganglia
- early growth response protein-4
- extracellular signal-regulated kinases-1/2
- total ERK
- E twenty six
- ETS-like transcription factor-1
- phosphorylated ELK-1
- FBJ murine osteosarcoma viral oncogene homolog-B
- firefly luciferase
- enhanced green fluorescent protein
- G protein-coupled receptor
- HEPES buffer saline
- human embryonic kidney
- K+/Cl− cotransporter-2
- protein kinase A
- protein kinase C
- pertussis toxin
- quantitative reverse transcriptase-polymerase chain reaction
- repressor element 1-silencing transcription
- sodium-activated sodium channel type VII α
- serum response element
- serum response factor
- signal transducers and activator of transcription
- ternary complex factor
- yellow fluorescent protein
- Received February 2, 2012.
- Accepted March 26, 2012.
- Copyright © 2012 by The American Society for Pharmacology and Experimental Therapeutics