Coupling of neuronal 5-HT7 receptors to activation of extracellular-regulated kinase through a protein kinase A-independent pathway that can utilize Epac

Authors

  • Stanley L. Lin,

    1. Department of Psychiatry, University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School, Piscataway, New Jersey, USA
    Search for more papers by this author
  • Nadine N. Johnson-Farley,

    1. Department of Psychiatry, University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School, Piscataway, New Jersey, USA
    Search for more papers by this author
  • David R. Lubinsky,

    1. Department of Psychiatry, University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School, Piscataway, New Jersey, USA
    Search for more papers by this author
  • Daniel S. Cowen

    1. Department of Psychiatry, University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School, Piscataway, New Jersey, USA
    Search for more papers by this author

Address correspondence and reprint requests to Dr Daniel S. Cowen, Department of Psychiatry, University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School, 125 Paterson Street, Suite 2200, New Brunswick, NJ 08901, USA. E-mail: cowends@umdnj.edu

Abstract

The roles of 3′,5′-cyclic adenosine monophosphate (cAMP) and protein kinase A in 5-hydroxytryptamine (5-HT)7 receptor-mediated activation of extracellular-regulated kinase (ERK) were studied in cultured hippocampal neurons and transfected PC12 cells. Activation of ERK by neuronal Gs-coupled receptors has been thought to proceed through a protein kinase A-dependent pathway. In fact we identified coupling of 5-HT7 receptors to activation of adenylyl cyclase and protein kinase A. However, no inhibition of agonist-stimulated ERK activation was found when cells were treated with H-89 and KT5720 at concentrations sufficient to completely inhibit activation of protein kinase A. However, activation of ERK was found to be sensitive to the adenylyl cyclase inhibitor 9-(tetrahydrofuryl)-adenine, suggesting a possible role for a cAMP-guanine nucleotide exchange factor (cAMP-GEF). Co-treatment of cells with 8-(4-chlorophenylthio)-2′-O-methyladenosine 3′,5′-cyclic monophosphate, a direct activator of the cAMP-GEFs Epac1 and 2, reversed the inhibition of agonist-stimulated ERK activation induced by adenylyl cyclase inhibition. Additionally, over-expression of Epac1 enhanced 5-HT7 receptor-mediated activation of ERK. These results demonstrate that the activation of ERK mediated by neuronal Gs-coupled receptors can proceed through cAMP-dependent pathways that utilize cAMP-GEFs rather than protein kinase A.

Abbreviations used
[Ca2+]

calcium concentration

cAMP

3′,5′-cyclic adenosine monophosphate

8-CPT-Me-cAMP

8-(4-chlorophenylthio)-2′-O-methyladenosine 3′,5′-cyclic monophosphate

5-CT

5-carboxamidotryptamine maleate

ERK

extracellular signal-regulated kinase

GEF

guanine nucleotide exchange factor

HA

Hemagglutinin

HEK

human embryonic kidney

5-HT

5-hydroxytryptamine

MAP

mitogen-activated protein

pCPT-cAMP

8-(4-chlorophenylthio)-3′,5′-adenosine cyclic monophosphate

PKA

protein kinase A

p-MPPI

4-iodo-N-[2-[4-(methoxyphenyl)-1-piperazinyl]ethyl]-N-2-pyridinyl-benzamide hydrochloride

Rp-8-Br-cAMPs

adenosine 3′,5′-cyclic monophosphorothioate, 8-bromo-, Rp-isomer

The selective serotonin reuptake inhibitors represent the most commonly prescribed class of antidepressants. Although relatively little is currently known about their mechanism of action, it is clear that the hippocampus is a target. Chronic treatment has been shown to cause cellular changes that would appear to be protective of stress- and glucocorticoid-induced damage. For example, psychological stress has been found to induce inhibition of hippocampal neurogenesis, an effect postulated to play a role in causing depression (Jacobs et al. 2000). Conversely, chronic treatment with selective serotonin reuptake inhibitors has been shown to stimulate hippocampal neurogenesis (Jacobs et al. 2000; Malberg et al. 2000) and to increase dendritic branching in the hippocampus (Duman et al. 1997).

The signaling pathways required for conferring these 5-hydroxytryptamine (5-HT)-induced neuroprotective changes are not currently known. However, a role for cellular pathways, such as extracellular signal-regulated kinase (ERK) mitogen-activated protein (MAP) kinase, that elicit long-term neuronal changes appears likely. The ERK pathway is known to enhance cell survival, and is required for normal neuronal functioning (Encinas et al. 1999; Erhardt et al. 1999). It is therefore intriguing that we have recently demonstrated that 5-HT stimulates activation of ERK in cultured rat (embryonic day 18) hippocampal neurons. Most, if not all, of this activation of ERK was found to be mediated through 5-HT7 receptors (Errico et al. 2001), which couple to the G protein Gs (Adham et al. 1998). In the present studies we have examined the pathway required for coupling these receptors to activation of ERK.

Agonists for Gs-coupled receptors cause inhibition of ERK activation in most cell types. This inhibition is thought to result, in part, from 3′,5′-cyclic adenosine monophosphate (cAMP)-dependent activation of protein kinase A (PKA), which has been shown to phosphorylate, and thereby inhibit, Raf-1 (Cook and McCormick 1993; Wu et al. 1993). However, ERK has been shown to be activated by agents that increase the level of cAMP in several cell types including neurons and cells of neuronal lineage, such as PC12 cells. A pathway for such activation has been proposed by Vossler et al. (1997) to require cAMP, PKA, Rap1, B-Raf and MAP kinase kinase (MEK). However, a number of variations on this pathway have also been reported, including a role for Ras and Raf-1 (Peraldi et al. 1995; Yamamori et al. 1995; Ambrosini et al. 2000; Norum et al. 2003).

Until recently, it was assumed that all cellular changes induced by cAMP are mediated by PKA. However, it is now known that increases in cAMP, such as those induced by Gs-coupled receptors, can also elicit effects independent of PKA. In many of these cases the effects are mediated by cAMP-activated guanine-nucleotide exchange factors (cAMP-GEFs or Epacs) (de Rooij et al. 1998; Kawasaki et al. 1998; Pham et al. 2000; Kashima et al. 2001; Laroche-Joubert et al. 2002; Mei et al. 2002). In light of these new findings, it is clear that the role of PKA needs to be directly examined in pathways previously assumed to be mediated by the kinase. Interestingly, cAMP-GEFs have been reported recently to stimulate activation of ERK in several non-neuronal cell types including melanocytes (Busca et al. 2000) and kidney cells (Laroche-Joubert et al. 2002). Because Epac 1 and Epac 2, cAMP-GEFs for Rap1 and Rap2, are expressed in brain (Kawasaki et al. 1998), these exchange factors may have similar roles in neurons. Indeed, the present findings support a role for one or more cAMP-GEFs in the pathway for coupling of 5-HT7 receptors to activation of ERK in neurons.

Experimental procedures

Materials

R -(+)-alpha-(2,3-dimethoxyphenyl)-1-[2-(4-fluorophenyl)ethyl]-4-piperidinemethanol (MDL 100907) was kindly provided by Aventis Pharmaceuticals (Bridgewater, NJ, USA). 9-(Tetrahydrofuryl)-adenine (SQ22 536), KT5720, and adenosine 3′,5′-cyclic monophosphorothioate, 8-bromo-, Rp -isomer (Rp-8-Br-cAMPs) were obtained from Calbiochem (San Diego, CA, USA). Thapsigargin was obtained from Calbiochem and from Alomone Laboratories (Jerusalem, Israel). 5-HT, N -[2-( p -bromocinnamylamino)ethyl]-5-isoquinolinesulfonamide (H-89), 8-(4-chlorophenylthio)-3′,5′-adenosine cyclic monophosphate (pCPT-cAMP), [ R ]-3-[2-(2-[4-methyl-piperidin-1-yl]ethyl)pyrrolidine-1-sulfonyl]phenol (SB269970), 4-iodo- N -[2-[4-(methoxyphenyl)-1-piperazinyl]ethyl]- N -2-pyridinyl-benzamide hydrochloride (p-MPPI), N -[2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl]- N -2-pyridinyl-cyclohexanecarboxamide maleate (WAY-100635) and 5-carboxamidotryptamine maleate (5-CT) were obtained from Sigma (St Louis, MO, USA). 8-(4-Chlorophenylthio)-2′- O -methyladenosine 3′,5′-cyclic monophosphate (8-CPT-Me-cAMP) was obtained from Axxora LLC (San Diego, CA, USA).

Cell culture

Hippocampal neuronal cultures were prepared as described previously (Levine et al. 1995; Errico et al. 2001). Hippocampi were isolated from embryonic day 18 Sprague–Dawley rats obtained from Hilltop Laboratory Animals (Scottsdale, PA, USA), and 106 cells were plated per poly-d-lysine-coated 35-mm petri dish. Cells were maintained in serum-free medium consisting of a 1 : 1 (v/v) mixture of Ham's F-12 and Eagle's minimal essential medium supplemented with 25 mg/mL insulin, 100 mg/mL transferrin, 60 mm putrescine, 20 nm progesterone, 30 nm selenium, 6 mg/mL glucose, 7.5 U penicillin and 7.5 µg streptomycin per mL at 37°C (95% air, 5% CO2). Cells remained in culture for 15 days before use.

PC12 cell culture

PC12 cells were obtained from American Type Culture Collection (Rockville, MD, USA), and were routinely cultured in Dulbecco's Modified Eagle's medium supplemented with l-glutamine, minimal essential medium non-essential amino acids, 15% dialyzed fetal bovine serum (dialyzed in membranes with 1000-Da molecular weight cut-offs against a 100-fold greater volume of 150 mm NaCl to remove endogenous 5-HT), 100 U penicillin and 100 µg streptomycin per mL at 37°C (95% air, 5% CO2). A stable, tightly adherent cell population was obtained after several cycles of washing off loosely adherent cells (Quinn et al. 2002). The gene for the human 5-HT7A receptor was isolated by PCR from Quick Clone human brain cDNA (Clonetech, Palo Alto, CA, USA). The gene was then cloned into a murine leukemia retroviral packaging vector and transduced into PC12 cells. Cells stably expressing receptors were selected in the presence of 500 µg/mL geneticin. Cells represented mixed cultures, as individual clones were not selected. This avoided the potential problem of selecting clones not representative of PC12 cells. The day before use, PC12 cells (but not hippocampal neurons which are already cultured under serum-free conditions) were washed with phosphate-buffered saline and cultured overnight under low-serum (0.5%) conditions.

Transient transfections of hemagglutinin

A cDNA encoding HA epitope-tagged Epac1 was kindly provided by Dr Johannes L. Bos. Transient transfections of PC12 cells with 9 µg plasmid cDNA per 60-mm cell culture dish were performed 48 h before cellular studies with Lipofectamine 2000 according to the manufacturer's suggestions (Gibco BRL, Rockville, MD, USA).

Immunoblots

Monoclonal anti-phospho-ERK1/ERK2 (Thr202/Tyr204) was obtained from Cell Signaling (Beverly, MA, USA). Rabbit polyclonal anti-total ERK1/ERK2 anti-HA epitope, and horseradish peroxidase-conjugated secondary antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). PC12 cells and hippocampal neurons were treated in the absence of phosphodiesterase and phosphatase inhibitors with the addition of reagents directly to the culture media. Cells were then washed three times with phosphate-buffered saline before lysis with a 26-gauge needle in 25 mm HEPES (pH 7.4), 150 mm NaCl, 1% Triton X-100, 1 mmβ-glycerol phosphate, 50 mm NaF, 5 mm EDTA, 1 mm sodium orthovanadate, 250 µm 4-(2-aminoethyl)-benzene-sulfonylfluoride hydrochloride, 0.1% aprotinin and 10 µg/mL leupeptin. Proteins were separated on 10% resolving gels (Bio-Rad Laboratories, Hercules, CA, USA) and transferred to 0.45-µm Immobilon-P polyvinylidene difluoride membranes (Millipore Corporation, Bedford, MA, USA). Bound antibodies were visualized using Enhanced Luminol Chemiluminescence Reagent (PerkinElmer Life Sciences, Boston, MA, USA) and direct exposure to a Kodak Image Station 440CF with a cooled, full-frame-capture CCD camera (Kodak, Rochester, NY, USA). Net intensity of bands was calculated directly from stored images using Kodak Digital Science 1D Image Analysis Software (version 3.5) on defined regions of interest. Data are expressed in figures as the means ± SEM (× 103) of three or more experiments performed in duplicate.

PKA assay

PC12 cells were treated in the absence of phosphodiesterase and phosphatase inhibitors with the addition of reagents directly to the culture media. Cells were then washed three times with phosphate-buffered saline before lysis with a 26-gauge needle in detergent-free lysis buffer [25 mm HEPES (pH 7.4), 50 mm NaF, 5 mm EDTA, 1 mm sodium orthovanadate, 250 µm 4-(2-aminoethyl)-benzene-sulfonylfluoride hydrochloride, 0.1% aprotinin and 10 µg/mL leupeptin]. The catalytic subunit of PKAα was immunoprecipitated during a 2-h incubation at 4°C with pre-complexed rabbit polyclonal anti-PKAα catalytic subunit and protein G PLUS-agarose, both obtained from Santa Cruz Biotechnology. Immunoprecipitated PKA-dependent incorporation of 32P into kemptide (Upstate, Lake Placid, NY, USA) was measured by scintillation spectrometry after spotting on P81 paper and extensive washing with 0.75% phosphoric acid and acetone. It should be noted that the kinase reaction was performed in the absence of cAMP in order to measure specifically the activity of PKA that had been activated by 5-CT before cell lysis. When experiments were performed to examine the in vitro inhibitory actions of KT5720, immunoprecipitated PKA was pretreated for 15 min at room temperature (22°C) with KT5720 before addition of Mg2+/ATP.

Measurement of cAMP

PC12 cells were treated in the absence of phosphatase inhibitors with the addition of reagents directly to the culture media. Before treatment with the other reagents, cells were pretreated for 5 min with 100 µm 3-isobutyl-I-methylxanthine, a phosphodiesterase inhibitor. Media was aspirated and discarded after treatment, and cells were lysed with a 26-gauge needle in 0.1 m HCl, without prior washing. Intracellular concentrations of cAMP were measured using a colorimetric cAMP direct immunoassay kit obtained from Calbiochem, according to the manufacturer's instructions.

Statistical analysis

Results are expressed as the mean ± SEM of three or more experiments, performed in duplicate. Experimental groups were compared by anova followed by Bonferroni post hoc tests.

Results

PC12 cells can be used as a model system for studying the coupling of 5-HT7 receptors to activation of ERK

In previous studies we have demonstrated that 5-HT7 receptors couple to activation of ERK in cultured hippocampal neurons (Errico et al. 2001). As part of our present efforts to delineate this receptor-mediated pathway, a transfected cell line stably expressing human 5-HT7A receptors was created. PC12 cells, commonly used in studies of neuronal signaling, were chosen for the parental cell line. Although these cells express endogenous Gq-coupled 5-HT2A receptors that couple to activation of ERK (Quinn et al. 2002), the pharmacology of 5-HT2A and 5-HT7 receptors is sufficiently different that agonists can be used to selectively activate one receptor without activating the other. Large increases in the level of activated, double-phosphorylated ERK1 and ERK2 were observed when cells stably expressing 5-HT7 receptors were treated with the 5-HT1/5-HT7 selective agonist 5-CT (Fig. 1a). Pretreatment with the selective 5-HT7 receptor antagonist SB269970 completely inhibited this activation, whereas pretreatment with 1 µm MDL 100907, a selective 5-HT2A receptor antagonist (Ki 0.9 nm for 5-HT2A; Kehne et al. 1996), caused no inhibition. In contrast, we have previously reported that as little as 10 nm MDL 100907 completely inhibits the activation of ERK stimulated by 5-HT in non-transfected PC12 cells through endogenous 5-HT2A receptors (Quinn et al. 2002). Additionally, we found that 5-CT, at concentrations of up to 1 µm, caused no activation of ERK in the parental cell line (not shown). Therefore, the 5-HT7 receptor-expressing PC12 cells represent a useful model system for selectively studying the coupling of 5-HT7 receptors to activation of ERK.

Figure 1.

5-CT stimulates activation of ERK in PC12 cells stably expressing 5-HT 7A receptors through a cAMP-dependent pathway. PC12 cells stably expressing human 5-HT 7A receptors were treated for 5 min with (a)  300 n m 5-CT in the absence or presence of 3 µ m SB269970 (SB) or 1 µ m MDL 100907 (MDL) and then lysed. (b)  PC12 cells stably expressing human 5-HT 7A were pretreated with 1 m m SQ22 536 (SQ) for 30 min before treatment with 30 n m 5-CT or 1 µ m forskolin (For) for 5 min. Supernatant was analyzed by immunoblotting with antibody to phospho-ERK1/ERK2 (p-ERK). Membranes were then stripped and analyzed with antibody to total ERK1/ERK2 (Total). Values are mean ± SEM. *** p  < 0.001 versus 5-CT or forskolin ( anova followed by Bonferroni post hoc test) . Representative immunoblots are also shown.

5-CT stimulates activation of ERK through a cAMP-dependent pathway

5-CT was found to stimulate a maximal 60-fold increase in intracellular cAMP concentration in PC12 cells expressing 5-HT7 receptors (Table 1). Direct activation of adenylyl cyclase with forskolin stimulated a magnitude of ERK activation and cAMP accumulation similar to that seen with 5-CT (Table 1 and Fig. 1b). In order to study the role of cAMP in 5-HT7 receptor-mediated ERK activation, we examined the effect of pretreating cells with the selective adenylyl cyclase inhibitor SQ22 536. SQ22 536 (1 mm) inhibited the activation of ERK stimulated by 5-CT by 50% (Fig. 1b). At this concentration SQ22 536 inhibited cAMP accumulation by more than 90% (Table 1). However, the remaining adenylyl cyclase activity still resulted in a fivefold increase in intracellular cAMP concentration. SQ22 536 inhibited forskolin-stimulated ERK activation and cAMP accumulation to an extent similar to that seen with 5-CT. We would have liked to test the effects of higher concentrations of SQ22 536 in order to determine whether more complete adenylyl cyclase inhibition would cause greater inhibition of ERK activation. However, this would have required exposing the cells to a larger volume of solvent (dimethyl sulfoxide). We have found that concentrations of dimethyl sulfoxide greater than the maximal 0.2% used in our studies independently inhibit 5-CT-stimulated activation of ERK (not shown).

Table 1.  Intracellular cAMP levels
 Intracellular cAMP (pmol)
  1. PC12 cells stably expressing human 5-HT7A were pretreated with 1 mm SQ22 536 for 30 min before treatment with 30 nm 5-CT or 1 µm forskolin for 5 min. ***P < 0.001 versus 5-CT or forskolin (anova followed by Bonferroni post hoc test).

Vehicle2.1 ± 0.5***
5-CT128 ± 27
SQ22 536 + 5-CT10.9 ± 2.1***
Forskolin167 ± 88
SQ22 536 + forskolin11.2 ± 4.0***

Activation of ERK is independent of PKA

Activation of ERK by neuronal Gs-coupled receptors has been thought to require PKA. We therefore developed an in vitro assay system to determine whether 5-CT stimulates activation of the kinase. PKA, in its inactive form, exists as a tetramer consisting of two catalytic subunits and two regulatory subunits. Upon binding cAMP, the regulatory subunits dissociate leaving free, active catalytic subunits. In order to assay cellular PKA activity, we immunoprecipitated the kinase with an antibody directed against the catalytic subunits and performed in vitro kinase assays in the absence of cAMP. Consequently, in vitro kinase activity represented the activity of catalytic subunits that had already been freed from regulatory subunits before cell lysis. We found that 5-CT stimulated an average threefold activation of PKA in PC12 cells expressing 5-HT7 receptors (Fig. 2a). We next used our assay system to determine the concentrations of H-89 and KT5720 required to inhibit activation of PKA. Because these are both reversible PKA inhibitors (Kase et al. 1987; Chijiwa et al. 1990), it was possible that they might dissociate from the kinase during the several hours between cell lysis and assay of PKA activity. However, we were able to demonstrate that pretreatment of cells with as little as 0.1 µm H-89 completely inhibited 5-CT-stimulated activation of PKA (Fig. 2a). In contrast, KT5720 apparently dissociated from the kinase during the time required for immunoprecipitation and assay of activity. PKA immunoprecipitated from cells pretreated with 1 µm KT5720 exhibited no decrease in activity compared with PKA derived from cells treated with 5-CT in the absence of KT5720. However, complete inhibition of 5-CT-stimulated protein kinase activity was observed when 1 µm KT5720 was included in the in vitro kinase assay (Fig. 2b). This demonstrated that 1 µm is sufficient to completely inhibit 5-CT-stimulated activation of PKA, the inhibition being reversible.

Figure 2.

0.1 µ m H-89 and 1 µ m KT5720 are sufficient to completely inhibit activation of protein kinase A. (a)  PC12 cells stably expressing human 5-HT 7A receptors were pretreated with the specified concentrations of H-89 for 30 min before treatment with 300 n m 5-CT for 5 min. PKA was then immunoprecipitated from the cell lysate. (b)  Cells were either pretreated for 30 min with 1 µ m KT5720 (KT) before 5-CT (PRE), or PKA immunoprecipitated from the lysate of cells treated solely with 5-CT was treated with 1 µ m KT5720 15 min before initiation of in vitro kinase assays (POST). The activity of immunoprecipitated PKA was assayed in the absence of cAMP in order to specifically measure the activity of catalytic subunits that had already been freed from regulatory subunits before cell lysis. Activity was calculated as microCuries of 32 P incorporated into kemptide per minute. Values are mean ± SEM. * p  < 0.05, ** p  < 0.01, *** p  < 0.001 versus 5-CT ( anova followed by Bonferroni post hoc test) .

In contrast, pretreatment of cells with 1 µm H-89 or 1 µm KT5720 caused no inhibition of 5-CT-stimulated ERK activation (Fig. 3). On the contrary, KT5720 potentiated the effects of 5-CT. Only when the concentration of H-89 was increased to 10 µm (100-fold higher than that required to inhibit PKA) did we observe a partial (50%) inhibition of ERK activation (Fig. 3a). K252a, an inhibitor of multiple kinases, including PKA, was found to be similar to KT5720 in inducing a twofold to threefold potentiation of 5-CT-stimulated activation of ERK (not shown). To further demonstrate that PKA was not required for 5-HT7 receptor-mediated activation of ERK, we tested the effects of a fourth inhibitor, Rp-8-Br-cAMPs. Unlike H-89, KT5720 and K252a, which cause inhibition by preventing binding of ATP to the kinase catalytic subunits, Rp-8-Br-cAMPs is an analog of cAMP that inhibits PKA activity by blocking binding of cAMP to the regulatory subunits. It was tested at concentrations of 200–500 µm, which have been shown to inhibit both type I and type II forms of PKA (Gjertsen et al. 1995). Even at the high concentration of 500 µm, Rp-8-Br-cAMPs had no inhibitory effect on 5-CT-mediated ERK activation (Fig. 3b).

Figure 3.

PKA inhibitors do not inhibit 5-CT-stimulated activation of ERK in PC12 cells at concentrations sufficient to inhibit activation of PKA. PC12 cells stably expressing human 5-HT 7A receptors were pretreated for 30 min with (a)  1 µ m H-89, 10 µ m H-89 or 1 µ m KT5720 (KT), or (b)  1 µ m H-89, 1 µ m KT5720 or 200–500 µ m Rp-8-Br-cAMPs (Rp) before treatment with 300 n m 5-CT for 5 min. Supernatant was analyzed by immunoblotting with antibody to phospho-ERK1/ERK2 (p-ERK). Membranes were then stripped and analyzed with antibody to total ERK1/ERK2 (Total). Values are mean ± SEM. * p  < 0.05, ** p  < 0.01, *** p  < 0.001 versus 5-CT ( anova followed by Bonferroni post hoc test) . Representative immunoblots are also shown.

Similar findings were obtained when the coupling of 5-HT7 receptors to activation of ERK was studied in cultured hippocampal neurons. A twofold increase in the level of activated, double-phosphorylated ERK was observed when hippocampal neurons were treated with 5-CT (Fig. 4a). Although both ERK1 (p44) and ERK2 (p42) were activated, most of the activation appeared to be of ERK2. The selective 5-HT7 receptor antagonist SB269970 completely inhibited this activation. In contrast, as we have previously reported, the selective 5-HT1A receptor antagonists WAY-100635 and p-MPPI caused no inhibition (Errico et al. 2001).

Figure 4.

PKA inhibitors do not inhibit activation of ERK in hippocampal neurons at concentrations sufficient to inhibit activation of PKA. (a)  Hippocampal neurons were treated for 5 min with 100 n m 5-CT in the absence or presence of 1 µ m SB269970 (SB), 1 µ m p-MPPI (MPPI) or 1 µ m WAY-100635 (WAY) and then lysed. (b)  Hippocampal neurons were pretreated with 0.1 µ m H-89 or 1 µ m KT5720 (KT) for 30 min before treatment with 1 µ m 5-CT for 5 min and then lysed. Supernatant was analyzed by immunoblotting with antibody to phospho-ERK1/ERK2 (p-ERK). Membranes were then stripped and analyzed with antibody to total ERK1/ERK2 (Total). Values are mean ± SEM. ** p  < 0.01, *** p  < 0.001 versus 5-CT ( anova followed by Bonferroni post hoc test) . Representative immunoblots are also shown.

Pretreatment of hippocampal cultures with 0.1 µm H-89 or 1 µm KT5720, concentrations shown in Fig. 2 to completely inhibit activation of PKA, induced no inhibition of 5-CT-stimulated ERK activation (Fig. 4b). Only at concentrations above that required to inhibit PKA was H-89 found to inhibit stimulation of ERK; 1 µm H-89 and 10 µm H-89 caused a 60% and 100% inhibition respectively of 5-CT-stimulated activity (Fig. 5a). In contrast, KT5720, even at the high concentration of 3 µm, caused no inhibition of ERK activation (not shown).

Figure 5.

Activation of ERK in hippocampal neurons is sensitive to high concentrations of H-89, but not to high concentrations of KT5720. Hippocampal neurons were pretreated with 1 µ m H-89 or 10 µ m H-89 for 30 min before treatment with (a) 1 µ m 5-CT for 5 min or (b) 10 µ m forskolin (For) for 10 min. Supernatant was analyzed by immunoblotting with antibody to phospho-ERK1/ERK2 (p-ERK). Membranes were then stripped and analyzed with antibody to total ERK1/ERK2 (Total). Values are mean ± SEM. ** p  < 0.01, *** p  < 0.001 versus 5-CT or forskolin ( anova followed by Bonferroni post hoc test) . Representative immunoblots are also shown.

Forskolin induced a twofold activation of ERK in cultured hippocampal neurons, a magnitude similar to that seen with 5-CT. Significantly, this cAMP-induced activation of ERK exhibited the same pattern of sensitivity to PKA inhibitors seen with 5-CT; 1 µm H-89 inhibited activation by 60% (Fig. 5b). Complete inhibition of forskolin-stimulated activation of ERK required 10 µm H-89, whereas 3 µm KT5720 caused a potentiation of ERK activation (not shown).

Activation of ERK can utilize Epac

It has been reported recently that ERK can be activated in melanocytes (Busca et al. 2000) and kidney cells (Laroche-Joubert et al. 2002) by cAMP-dependent pathways that involve cAMP-GEFs, instead of PKA. Treatment of PC12 cells with pCPT-cAMP, a cAMP analog that activates both PKA and Epac, was found to stimulate ERK activation (Fig. 6a). In contrast, 8-CPT-Me-cAMP, a recently described selective activator of Epacs 1 and 2 (Enserink et al. 2002), did not stimulate activation of ERK. This suggested that direct activation of endogenous Epac is not sufficient to induce ERK activation. However, Epac did appear to play a role in 5-HT7 receptor-mediated ERK activation. Co-treatment of cells with 8-CPT-Me-cAMP reversed the inhibition of 5-CT-stimulated ERK activation induced by the adenylyl cyclase inhibitor SQ22 536. We next used transfected PC12 cells to examine whether over-expression of Epac could potentiate the activation of ERK mediated by 5-HT7 receptors. Indeed, transient transfection with cDNA for Epac1 caused a potentiation of 5-CT-stimulated activation of endogenous ERK (Fig. 6b).

Figure 6.

5-HT 7 receptor-mediated activation of ERK can utilize Epac. (a)  PC12 cells stably expressing human 5-HT 7A were treated, as indicated, with 1 m m SQ22 536 (SQ) or vehicle for 30 min before treatment with 200 µ m pCPT-cAMP (pCPT) for 10 min, 100 µ m 8-CPT-Me-cAMP (8-CPT-Me) for 10 min, and/or 30 n m 5-CT for 5 min. (b)  PC12 cells stably expressing human 5-HT 7A were transiently transfected with cDNA for HA-tagged Epac1 (EPAC) or a control plasmid. Cells were then treated with 30 n m 5-CT for 5 min before lysis. Supernatant was analyzed by immunoblotting with antibody to phospho-ERK1/ERK2 (p-ERK). Membranes were then stripped and analyzed with antibody to total ERK1/ERK2 (Total) and the HA-tag. Values are mean ± SEM. (a) *** p  < 0.001 versus vehicle treated. (b) *** p  < 0.001 versus control plasmid-transfected cells treated with 5-CT ( anova followed by Bonferroni post hoc test) . Representative immunoblots are also shown.

5-HT7 receptor-mediated activation of ERK in PC12 cells does not require an increase in intracellular calcium concentration ([Ca2+])

Our finding that direct stimulation of endogenous Epacs with 8-CPT-Me-cAMP did not activate ERK suggested that ERK activation requires the concomitant activation of additional pathways. Interestingly, agonists for 5-HT7 receptors have been shown to stimulate increases in intracellular [Ca2+] in rat glomerulosa cells and transfected human embryonic kidney (HEK) 293 cells (Baker et al. 1998; Lenglet et al. 2002). We therefore examined the role of Ca2+ in the pathway for activation of ERK. No inhibition of PC12 ERK activation was found when extracellular [Ca2+] was reduced from 2 mm to 100 nm (a concentration similar to the intracellular [Ca2+] seen in many types of resting cells) by pretreatment with EGTA (Fig. 7a). On the contrary, reduction of extracellular [Ca2+] caused a potentiation of 5-CT-stimulated activity. Similarly, pretreatment of cells with 30 nm thapsigargin, to slowly deplete intracellular stores of Ca2+ before treatment with 5-CT, caused a potentiation, not inhibition, of ERK activation. In contrast, we have previously demonstrated that treatment with the same concentrations of EGTA and thapsigargin inhibit activation of ERK by endogenous Gq-coupled 5-HT2A receptors (Quinn et al. 2002).

Figure 7.

5-HT 7A receptor-mediated activation of ERK in PC12 cells and hippocampal neurons does not require increases in intracellular [Ca 2+ ]. (a)  PC12 cells stably expressing human 5-HT 7A receptors or (b)  hippocampal neurons were pretreated with EGTA (to reduce extracellular [Ca 2+ ] to 100 n m ) or 30 n m thapsigargin (Thap) for 30 min before treatment with 5-CT for 5 min. Supernatant was analyzed by immunoblotting with antibody to phospho-ERK1/ERK2 (p-ERK). Membranes were then stripped and analyzed with antibody to total ERK1/ERK2 (Total). Values are mean ± SEM. * p  < 0.05, ** p  < 0.01, *** p  < 0.001 versus 5-CT ( anova followed by Bonferroni post hoc test) . Representative immunoblots are also shown.

As was found in PC12 cells, 5-HT7 receptor-mediated activation of ERK in hippocampal neurons was independent of Ca2+. Neither pretreatment with EGTA nor thapsigargin caused inhibition of 5-CT-stimulated activation of ERK (Fig. 7b).

Discussion

With the discovery of the cAMP-GEFs, it is now clear that PKA plays a role in only a subset of the pathways regulated by Gs-coupled receptors. In fact, our studies demonstrate that increases in cAMP can cause activation of ERK in neurons through a pathway independent of PKA. In the course of these studies we found that the PKA inhibitor H-89 has cellular effects, independent of PKA, that interfere with cAMP-dependent activation of ERK. Although 0.1 µm H-89 was sufficient to completely inhibit 5-CT-stimulated activation of PKA, it did not inhibit activation of ERK. In contrast, higher concentrations of H-89 were found to inhibit activation of both PKA and ERK. This finding is significant in that the kinase inhibitor is commonly used in cellular studies at concentrations of 10 µm or higher (Chijiwa et al. 1990; Peraldi et al. 1995; Kiehn et al. 1998; Wang and Brown 1999; Grewal et al. 2000; Schmitt and Stork 2002). In fact, some studies have reported that 20–30 µm concentrations are required to elicit specific desired cellular effects (Chijiwa et al. 1990), although the reported Ki for inhibition of purified PKA is only 40 nm (Chijiwa et al. 1990). It has been hypothesized (but not demonstrated) that the cell permeability of H-89 may be limited, such that micromolar extracellular concentrations are required to achieve nanomolar intracellular concentrations. However, our studies demonstrate that this is not the case. H-89 at a concentration of 0.1 µm was sufficient to completely inhibit activation of PKA in intact cells.

Norum et al. (2003 ) recently characterized the coupling of 5-HT 7 receptors to activation of ERK in transfected HEK 293 cells. Activation of ERK was shown to require Ras and to be independent of Rap1. Similar to our findings, activation of ERK was found to be inhibited by high concentrations of H-89 (20 µ m ). The effect of lower concentrations of H-89 was not reported. Although we found no requirement for PKA in the activation of ERK in hippocampal neurons and PC12 cells, it is possible that PKA might be required in the non-neuronal HEK 293 cells. In fact, the neuronal pathway appears to be somewhat different to the HEK 293 pathway. Raf-1 was reported to be required for coupling of receptors to activation of ERK in HEK 293 cells. However, we have not found activation in kinase assays of immunoprecipitated Raf-1 from PC12 cells treated with 5-CT (not shown).

Although we found no role for PKA in the activation of ERK by 5-CT and forskolin, we did find a requirement for cAMP. Pretreatment of cells with 1 mm SQ22 536, a selective inhibitor of adenylyl cyclase, inhibited activation of ERK by 50%. The same concentration of SQ22 536 inhibited by 40% the activation of ERK induced by direct activation of adenylyl cyclase with forskolin. Although treatment with SQ22 536 inhibited adenylyl cyclase activity by more than 90%, the residual activity still resulted in a fivefold increase in intracellular cAMP concentration. We would have liked to determine whether more complete inhibition of adenylyl cyclase activity would have resulted in more complete inhibition of ERK activation. However, we were unable to test the effect of higher concentrations because it would have required treating the cells with larger volumes of the solvent dimethyl sulfoxide, which inhibits activation of ERK at concentrations above the maximal 0.2% used in our studies.

We had expected that 5-HT7 receptor-mediated activation of ERK might be Ca2+ dependent in addition to being cAMP dependent. Agonists for 5-HT7 receptors have been shown to stimulate increases in intracellular [Ca2+] in rat glomerulosa cells and transfected HEK 293 cells (Baker et al. 1998; Lenglet et al. 2002). However, we found that neither an influx of extracellular [Ca2+] nor release of intracellular [Ca2+] stores was required for 5-HT7 receptor-mediated activation of ERK in PC12 cells and hippocampal neurons. On the contrary, we found a potentiation of ERK activation in PC12 cells when intracellular [Ca2+] was clamped. This inhibitory effect of Ca2+ might be mediated by the type VI adenylyl cyclase, which is expressed in PC12 cells (Chern et al. 1995) and is known to be inhibited by Ca2+.

Our results are consistent with a model in which 5-HT7 receptor-mediated activation of ERK requires a cAMP-GEF, such as Epac. We found that activation of ERK was sensitive to inhibition of adenylyl cyclase but not PKA. Significantly, we found that the selective Epac activator 8-CPT-Me-cAMP reversed the inhibition of ERK induced by the adenylyl cyclase inhibitor SQ22 536. Additionally, over-expression of Epac1 was found to potentiate activation of ERK. A number of cAMP-GEFs are expressed in brain, including Epac1, Epac2 and CNrasGEF (de Rooij et al. 1998; Kawasaki et al. 1998; Ozaki et al. 2000; Pham et al. 2000; Fujita et al. 2002). PC12 cells, like brain, have been reported to express message for both Epac1 and Epac2 (Fujita et al. 2002). However, the level of Epac2 message expressed in various strains of PC12 cells may vary, as Ozaki et al. (2000) reported detecting none in their PC12 cultures. Although our studies demonstrated that the pathways required for coupling 5-HT7 receptors to activation of ERK can utilize Epac1, we cannot be sure that endogenous Epac1 is the cAMP-GEF responsible for activation of ERK in PC12 cells and hippocampal neurons. Because selective cAMP-GEF inhibitors are not currently available, we cannot determine the roles of specific cAMP-GEFs. As an analog of cAMP, it might be expected that Rp-8Br-cAMPs, used in our studies to inhibit PKA, would also be useful for inhibiting cAMP-GEFs. However, at least with respect to Epac, it does not appear to be a potent inhibitor (Kang et al. 2003).

Our finding that direct activation of Epac with 8-CPT-Me-cAMP did not stimulate activation of ERK suggests that an additional cell component or pathway is required for 5-HT7 receptor-mediated ERK activation. This conclusion is consistent with the report by Enserink et al. (2002) demonstrating that 8-CPT-Me-cAMP stimulates activation of Rap1 but not ERK in PC12 cells. It is possible that concomitant activation of another cAMP-GEF is required for ERK to be activated through Epac. Such a model is consistent with our finding that direct activation of adenylyl cyclase with forskolin was sufficient to activate ERK through a PKA-independent pathway. In summary, our findings demonstrate that 5-HT7 receptors couple to activation of ERK through a cAMP-dependent pathway that utilizes one or more cAMP-GEFs, rather than through an expected PKA-dependent pathway.

Acknowledgement

These studies were supported by National Institute of Mental Health (NIMH) grant MH60100 to DSC.

Ancillary