5-Hydroxytryptamine type 7 receptor neuroprotection against NMDA-induced excitotoxicity is PDGFβ receptor dependent

Authors


Abstract

The serotonin (5-HT) type 7 receptor is expressed throughout the CNS including the hippocampus. Long-term (2–24 h) activation of 5-HT7 receptors regulates growth factor receptor expression, including the expression of platelet-derived growth factor (PDGF) β receptors. Direct activation of PDGFβ receptors in primary hippocampal and cortical neurons inhibits NMDA receptor activity and attenuates NMDA receptor-induced neurotoxicity. Our objective was to investigate whether the 5-HT7 receptor-induced increase in PDGFβ receptor expression would be similarly neuroprotective. We demonstrate that 5-HT7 receptor agonist treatment in primary hippocampal neurons also increases the expression of phospholipase C (PLC) γ, a downstream effector of PDGFβ receptors associated with the inhibition of NMDA receptor activity. To determine if the up-regulation of PDGFβ receptors is neuroprotective, primary hippocampal neurons were incubated with the 5-HT7 receptor agonist, LP 12, for 24 h. Indeed, LP 12 treatment prevented NMDA-induced neurotoxicity and this effect was dependent on PDGFβ receptor kinase activity. Treatment of primary neurons with LP 12 also differentially altered NMDA receptor subunit expression, reducing the expression of NR1 and NR2B, but not NR2A. These findings demonstrate the potential for providing growth factor receptor-dependent neuroprotective effects using small-molecule ligands of G protein-coupled receptors.

Abbreviations used
5-HT

5-hydroxytryptamine or serotonin

CNS

central nervous system

EPSPs

excitatory post-synaptic potentials

GPCR

G protein-coupled receptor

LTP

long-term potentiation

NMDA

N-methyl-D-aspartate

PDGF

platelet-derived growth factor

PLC

phospholipase C

SCN

suprachiasmatic nucleus

The serotonin (5-HT) 7 receptor was the last of the 5-HT receptors to be cloned (Hoyer and Martin 1997; Hoyer 2002; Kroeze et al. 2002). 5-HT receptors are G protein-coupled receptors with the exception of the type 3 5-HT receptor, which is a ligand-gated ion channel (van Hooft and Vijverberg 2000). 5-HT7 receptors are expressed in the hippocampus and are also expressed in the suprachiasmatic nucleus, pre-frontal cortex, the thalamus, and the hypothalamus (Hirst et al. 1997; Stam et al. 1997; Mahe et al. 2004; Thomas and Hagan 2004). Activation of 5-HT7 receptors leads to an accumulation of cyclic AMP in several systems (Thomas et al. 1999; Krobert et al. 2001; Crider et al. 2003; Mahe et al. 2004; Andressen et al. 2006; Romero et al. 2006); however, there is evidence that 5-HT7 receptors can couple to other G proteins. In NIH3T3 cells and organotypic hippocampal slices, 5-HT7 receptors are reported to couple to Gα12 and activate monomeric G proteins including Cdc42 and RhoA (Kvachnina et al. 2005; Kobe et al. 2012).

Clinically used antipsychotic drugs antagonize 5-HT7 receptors and compounds with 5-HT7 receptor antagonist activity have been investigated as potential treatments for schizophrenia and other psychoses (Uzun et al. 2005; Galici et al. 2008), and clinical trials are ongoing with novel 5-HT7 receptor antagonists, JNJ-18038683 and Lu AA21004, to treat depression (Bonaventure et al. 2012; Mork et al. 2012). Despite intense clinical research into drugs that activate or antagonize 5-HT7 receptors, very little is known about their molecular pharmacology, particularly their signaling roles in neurons. With respect to neuronal activity, 5-HT7 receptors are reported to increase excitability in the hippocampus, possibly by decreasing slow afterhyperpolarizations (Bacon and Beck 2000; Tokarski et al. 2005) or by inhibiting calcium-dependent potassium channels (Gill et al. 2002). 5-HT7 receptors have also been linked to the stimulation of afterdepolarizations in the thalamus (Chapin and Andrade 2001a,b). Thus, there is evidence that 5-HT7 receptors can significantly affect neuronal activity, but the molecular mechanisms remain unknown.

We have demonstrated that 2- to 24-h 5-HT7 receptor agonist treatment increases the expression of platelet-derived growth factor (PDGF)β receptors in primary hippocampal and cortical neurons, as well as in the SH-SY5Y cell line (Vasefi et al. 2012). The up-regulated PDGFβ receptors displayed an increase in basal phosphorylation at tyrosine 1021, the PLCγ docking site, and the phosphorylation site associated with PDGF-induced inhibition of NMDA-evoked currents (Valenzuela et al. 1996; Beazely et al. 2006; Vasefi et al. 2012). Both the expression of PDGFβ receptors and the increase in their phosphorylation level were maximally increased after a 24-h incubation, but increases were observed as early as 2 h (Vasefi et al. 2012).

The expression of the PDGF ligands, PDGF-A & -B mRNA and the PDGF-BB & -AB protein levels, is increased in neurons and support cells surrounding areas damaged during ischemic events (Krupinski et al. 1997). Exogenously applied PDGF-BB is also neuroprotective in focal ischemia (Sakata et al. 1998) and against glutamate- and NMDA-induced cell death in hippocampal neurons (Egawa-Tsuzuki et al. 2004; Tseng and Dichter 2005). We have demonstrated previously that PDGFβ receptors selectively inhibit NR2B-containing NMDA receptor currents and that PDGFβ receptor-mediated neuroprotection against NMDA-induced toxicity is associated with the inhibition of NR2B-containing NMDA receptors (Beazely et al. 2009). Acute treatment of hippocampal slices with PDGF-BB also selectively reduces the cell-surface localization of the NR1 and NR2B NMDA receptor subunits, but not NR2A (Beazely et al. 2009).

There are several challenges involved in the clinical use of growth factors to treat CNS disorders. Growth factors are proteins that are relatively large (compared with small-molecule drugs), which prevents both oral administration and their ability to passively cross the blood–brain barrier. Given the ability of 5-HT7 receptor agonists to up-regulate the expression and basal phosphorylation of PDGFβ receptors, our objective was to determine if simply up-regulating PDGFβ receptors would result in neuroprotective effects against NMDA-induced toxicity. We demonstrate that the 5-HT7 agonist, LP 12, increased PDGFβ receptor expression on pyramidal neurons in primary hippocampal cultures, and LP 12 treatment for 24 h prevented NMDA-induced cell death in a PDGFβ-receptor-dependent manner. Interestingly, 5-HT7 receptor activation decreased the expression of NR1 and NR2B NMDA receptor subunits but not the expression of NR2A. Thus, 5-HT7 receptor activation reduces NR2B expression and increases the expression of the PDGFβ receptor, a selective inhibitor of NR2B-containing NMDA receptors.

Materials and methods

Reagents and antibodies

LP 12 (4-(2-Diphenyl)-N-(1,2,3,4-tetrahydronaphthalen-1-yl)-1- piperazinehexanamide hydrochloride), NMDA, PDGF-BB, glycine, and other chemical reagents were purchased from Sigma (St. Louis, MO, USA). AG1296 (6,7-Dimethoxy-2-phenylquinoxaline) and Ro 25-6981 ((αR,βS)-α-(4-Hydroxyphenyl)-β-methyl-4-(phenylmethyl)-1-piperidinepropanol maleate) were purchased from Cedarlane (Burlington, ON, Canada). Imatinib mesylate (STI-571) (4-[(4-methyl-1-piperazinyl)methyl]-N-[4-methyl-3-[[4-(3-pyridinyl)-2-pyrimidinyl]amino]phenyl]-methanesulfonate-benzamide) was purchased from Novartis (Basel, Switzerland). The 5-HT7 receptor antagonists SB 258719 ((R)-3,N-Dimethyl-N-[1-methyl-3-(4-methylpiperidin-1-yl)propyl]benzene sulfonamide) and SB 269970 ((2R)-1-[(3-Hydroxyphenyl)sulfonyl]-2-[2-(4-methyl-1-piperidinyl)ethyl] pyrrolidine hydrochloride) were obtained from Tocris (Ellisville, MO, USA). Antibodies used include those raised against PDGFβ receptor (Epitomics, Burlinghame, CA, USA), PDGFβ receptor phospho-tyrosine 1021, PLCγ, β-actin (Santa Cruz, CA, USA), anti-NR1, anti-NR2A, and anti-NR2B (EMD Millipore, Billerica, MA, USA). All secondary antibodies including Dylight 488 were obtained from Fisher (Ottawa, ON, Canada).

Primary cell culture

Hippocampal neurons were isolated from E17-19 CD1 mouse fetuses (Harlan, Indianapolis, IN) and placed in cold dissection media (33 mM glucose, 58 mM sucrose, 30 mM HEPES, 5.4 mM KCl, 0.44 mM KH2PO4, 137 mM NaCl, 0.34 mM Na2HPO4, 4.2 mM NaHCO3, 0.03 mM phenol red, pH 7.4, 320–335 mOsm/kg). The hippocampi were separated from the brain and digested in 0.25% trypsin/0.1% EDTA for 20 min at 37°C. The resulting cell mixtures were plated on poly-d-lysine (PDL) (Sigma)-coated culture dishes and grown at 37°C in a humidified atmosphere containing 5% CO2. Cells were plated in plating media [Dulbecco's modified Eagle's medium (Fisher), 10% fetal bovine serum and 10% horse serum (Sigma)] and fed twice per week with feeding media [Neurobasal media supplemented with B27 (Life Technologies, Burlington, ON, Canada)]. The mitotic inhibitor FUDR (0.081 mM 5-fluoro-2-deoxyuridine) and 0.2 mM uridine (Sigma) were added 3–5 days after plating for 24 h once the cells reached confluency to inhibit the growth of non-neuronal cells. Drug treatments were performed 10–14 days after plating.

Western blot

After drug treatment, cells were washed with chilled phosphate-buffered saline (PBS), and lysed in chilled lysis buffer [20 mM Tris-HCl at pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 30 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM sodium orthovanadate, and 1% Triton X-100; supplemented with Halt Protease and Phosphatase Inhibitor (Thermo, Fisher, Markham, ON, Canada)] prior to use. For NMDA receptor subunit blotting, lysis buffer was supplemented with 1% sodium dodecyl sulfate and 1% Triton X-100 to solubilize the receptors. Cells were scraped, homogenized, and centrifuged at 14 000 g for 20 min at 4°C and the supernatant was collected. Homogenates were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and proteins were transferred to nitrocellulose membranes, blocked with 5% non-fat dry milk in Tris-buffered saline and 0.1% Tween-20 for 1 h at 22°C or overnight at 4°C, and incubated in primary antibodies for 1 h at 22°C or overnight at 4°C. Membranes were washed three times in Tris-buffered saline with 0.1% Tween-20, incubated with horseradish peroxidase-conjugated secondary antibodies for 1 h at 22°C, washed again, and bound antibodies were visualized by the enhanced chemiluminescence using Luminata Crescendo substrate (Millipore, Etobicoke, ON, Canada). Images of western blots were taken using a Kodak 4000 MM Pro Imaging Station, and densitometric analyses were performed using Kodak Molecular Imaging software (Carestream, Rochester, NY, USA). Membranes were then stripped and reprobed with additional antibodies.

Immunofluorescence

Primary hippocampal neurons were grown on sterile coverslips in Petri dishes and treated with LP 12 for 24 h. Hippocampal neurons were fixed using 4% (w/v) paraformaldehyde (Sigma) for 15 min and rinsed three times with PBS. The cultures were permeabilized using 0.3% Triton X-100 (Sigma) for 15 min and then washed with PBS. Hippocampal neurons were blocked in 4% (w/v) bovine serum albumin (Rockland, Gilbertsville, PA, USA) for 1 h and incubated with primary antibody overnight at 4°C with agitation. The coverslips were washed three times with PBS, labeled with secondary antibody (Dylight 488) at 1 : 2000 (Thermo) for 1 h, mounted with Prolong Gold antifade solution with DAPI (Life Technologies, Inc., Burlington, ON, Canada), and visualized using Zeiss Axiovert 200 microscope and LSM 510 META software (Carl Zeiss Canada Ltd., Mississauga, ON, Canada) with 63x oil-based objective.

MAP2 cell death assay

Hippocampal neurons, 10–14 days old, on 12-well plates, were pre-treated for 24 h with drugs, washed, and incubated with vehicle or NMDA/glycine for 5 min. The cells were washed and returned to culture medium overnight. After 24 h, cells were fixed in 4% paraformaldehyde in PBS, permeabilized with 0.4% Triton in PBS, blocked for 30 min with 10% bovine serum albumin in PBS, and incubated with anti-MAP2 antibodies (1 : 5000, Sigma) overnight at 4°C. The following day, the cells were washed in 0.4% Triton/PBS three times and incubated with anti-mouse IgG horseradish-peroxidase-conjugated antibody (Life Technologies, Inc., Burlington, ON, Canada). Cells were washed three times and incubated with Amplex UltraRed substrate for 30 min. The reaction was then terminated using Amplex UltraRed Stop Solution. Absorption at 565 nm was measured and data were normalized to control samples.

MTT toxicity assay

Hippocampal neurons were plated at density of 4 × 105 per mL in 96-well plates and grown for 10–14 days. The cells were pre-treated with drugs for 24 h and exposed to NMDA/glycine for 5 min. Cells were washed and incubated in feeding media for 24 h. 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) in an amount equal to 10% of the culture medium volume was added to each well and the cells were incubated for an additional 3 h at 37°C in 5% CO2. After incubation, the resulting formazan crystals were solubilized with MTT reagent solution (10% Triton X-100 and 0.1 N HCl in anhydrous isopropanol) in each well and the absorbance was recorded at 570 nm. All results were expressed as a percent reduction of MTT relative to untreated controls.

Statistical analysis

Statistical analysis of the data was completed using Prism® GraphPad program (Graphpad Software, La Jolla, CA, USA). All data are reported as mean ± SEM. Significance level is set at α = 0.05. Data were analyzed by one-way anova.

Animals

All animal experiments were performed in agreement with the guidelines of the policies on the Use of Animal at the University of Waterloo. All animal protocols were approved by the University of Waterloo animal care committee.

Results

We recently reported that long-term (2–24 h) activation of 5-HT7 receptors with 5-HT, 5-carboxamidotryptamine, or the 5-HT7 receptor-selective agonist, LP 12 (Leopoldo et al. 2008), increased PDGFβ receptor expression in primary mouse hippocampal and cortical neurons and in SH-SY5Y cells (Vasefi et al. 2012). The increase in PDGFβ receptor expression after LP 12 incubation was dose dependent with significant increases observed at concentrations of 100 and 300 nM (Fig. 1a, d). The increased expression of PDGFβ receptors was accompanied by an increase in the basal PDGFβ receptor phosphorylation state at tyrosine 1021 (Vasefi et al. 2012) as well as other tyrosine residues (data not shown). Tyrosine 1021 phosphorylation was maximally increased at 300 nM LP 12 (Fig. 1b, d). Similar to other Gαs-coupled receptors, 5-HT7 receptors are reported to be internalized and down-regulate after exposure to agonist (Krobert et al. 2006). Interestingly, despite the fact that LP 12 has a reported affinity of 0.13 nM for the 5-HT7 receptor, we only observed a significant reduction in 5-HT7 receptor expression after 24 h treatment with 300 nM LP 12 (Fig. 1c, d), the same concentration that maximally increased the expression of PDGFβ receptors. LP 12 also increased the expression of phospholipase C (PLC) γ in primary hippocampal cultures and this increase was blocked by the 5-HT7 receptor antagonist SB 258719 (1 μM, Fig. 1e, f).

Figure 1.

Effect of LP 12 on platelet-derived growth factor (PDGF)β and serotonin (5-HT)7 receptor and PLCγ expression in primary hippocampal neurons. (a) Neurons were treated with 10 nM–1 μM LP 12 for 24 h. The level of PDGFβ receptor expression is displayed as a fold change versus vehicle-treated neurons. Immunoreactivity for the PDGFβ receptor was normalized to β-actin. = 3 *< 0.05, **< 0.01, anova analysis with Dunnett's post-test. (b) Neurons were treated as in (a). The phosphorylation of the PDGFβ receptors was determined using an anti-phospho-Y1021 PDGFβ receptor antibody and was normalized to total PDFGβ receptor expression. = 3 *p < 0.05, anova analysis with Dunnett's post-test. (c) Neurons were treated as in A and B. Immunoreactivity for the 5-HT7 receptor was normalized to β-actin. = 4 ** < 0.01, anova analysis with Dunnett's post-test. (d) Representative western blots for the PDGFβ receptor, phospho-Y1021, the 5-HT7 receptor, and β-actin. (e) Hippocampal cultures were treated with vehicle or 300 nM LP 12, 1 μM SB 258719, or both for 24 h. The level of PLCγ expression is expressed as the fold change versus vehicle-treated cells. Immunoreactivity for PLCγ was normalized to β-actin, = 3 *< 0.05 vehicle versus LP 12, #< 0.05 LP 12 versus LP 12 + SB 258719, anova analysis with Bonferroni's post-test. Representative western blots are shown in (f).

There are two PDGF receptor isoforms, α and β, and in the hippocampus, the β isoform appears to be primarily expressed in pyramidal neurons (Smits et al. 1991; Beazely et al. 2009). PDGFβ receptors do not appear to significantly colocalize with PSD-95 or the NR2A subunit of the NMDA receptor, but have a relatively higher colocalization with the NR2B subunit (Beazely et al. 2009). To qualitatively assess if PDGFβ receptors after 24 h LP 12 treatment displayed a similar expression pattern compared with untreated cells, primary hippocampal neurons were treated for 24 h with vehicle or with 300 nM LP 12, fixed, and incubated with anti-PDGFβ receptor antibodies. After LP 12 treatment, PDGFβ receptor expression was still observed primarily on pyramidal neurons compared with non-neuronal cells (Fig. 2). Similar to our previous results, PDGFβ receptors were expressed on both neuronal cell bodies and processes.

Figure 2.

LP 12 increases platelet-derived growth factor (PDGF)β receptor expression in pyramidal neurons in primary hippocampal cultures. Cells were cultured on glass coverslips and maintained at 37°C for 14–21 days in vitro, at which time cells were treated with the control (vehicle) or with 300 nM LP 12 for 24 h. Nuclei were directly stained with DAPI (blue), and the PDGF receptors were detected with an anti-PDGF receptor antibody and secondary antibody conjugated to Dylight 488 (green). For the LP 12-treated images, the red arrow indicates a non-neuronal cell, which shows a DAPI-stained nucleus without PDGFβ receptor immunoreactivity. The images displayed are representative of six independent experiments.

Direct activation of PDGFβ receptors by PDGF-BB produces neuroprotective effects in cultured hippocampal neurons against glutamate- and NMDA-induced neurotoxicity (Tseng and Dichter 2005; Beazely et al. 2009). To determine if 24 h treatment with LP 12 could similarly promote neuroprotective effects by up-regulating PDGFβ receptor expression, we incubated primary hippocampal neurons for 24 h with 300 nM LP 12 followed by 100 μM NMDA/1 μM glycine for 10 min. We measured the remaining number of live cells using both MTT (Fig. 3a) and MAP2 (Fig. 3b) assays (Carrier et al. 2006; Beazely et al. 2009) 24 h post-NMDA treatment. As a positive control, we also treated neurons with PDGF-BB (10 ng/mL) for 10 min to directly activate PDGFβ receptors prior to NMDA treatment (Fig. 3a). LP 12 treatment for 24 h was indeed able to prevent NMDA-induced cell death (Fig. 3a, b). To determine if the neuroprotective effects of LP 12 required PDGFβ receptor activity, we repeated the cell death assay in the absence or presence of PDGFβ receptor kinase inhibitors. STI-571 inhibits PDGFβ receptor kinase activity in vitro with an IC50 value of ~100 nM (Roussidis et al. 2007), and inhibits PDGFβ receptor signaling with an IC50 value in the mid-μM range (Roussidis et al. 2007; Weigel et al. 2009). STI-571 attenuated the neuroprotective effects of LP 12 at concentrations of 1 and 5 μM (Fig 3c). Similarly, AG1296 prevented LP 12-induced cell death at 5 μM, a concentration similar to its IC50 for inhibiting PDGF-induced cell proliferation (Kovalenko et al. 1994) (Fig. 3d). Both STI-571 and AG1296 blocked the neuroprotective effects elicited by LP 12, suggesting that the neuroprotective effects are PDGFβ receptor dependent. The neuroprotective effects elicited by LP 12 were similar to those observed in neurons pre-treated with the NR2B-subunit-selective antagonist, Ro 25-6981 (data not shown, Beazely et al. 2009).

Figure 3.

Activation of serotonin (5-HT)7 receptors is neuroprotective via platelet-derived growth factor (PDGF)β receptor activity. Hippocampal cultures were pre-treated with vehicle or 300 nM LP 12 for 24 h followed by 100 μM NMDA/1 μM glycine for 10 min. For the PDGF-BB bar, 10 ng/mL PDGF-BB was added for 10 min prior to NMDA treatment. The number of cells was determined 24 h later by both 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (a, = 8) and MAP2 (b, = 7) assays. For (a), ***p < 0.001, vehicle versus NMDA, # p < 0.05, NMDA versus NMDA + LP, anova analysis with Bonferroni's post-test (For PDGF-BB, the last bar in (a) is a single, representative experimental result). For (b), *< 0.05, vehicle versus NMDA, #, < 0.01, NMDA versus NMDA + LP, anova analysis with Bonferroni's post-test. To determine if the neuroprotective effects of the 5-HT7-receptor agonist required PDGFβ receptor kinase activity, cultures were co-pre-treated with 1 or 5 μM STI-571 (c, = 13) or 1, 2.5, or 5 μM AG1296 (d, = 4) and 300 nM LP 12 for 24 h followed by 100 μM NMDA/1 μM glycine for 10 min. The number of cells was determined 24 h later by MTT assay. For (c and d), *p < 0.05, **< 0.01 compared with vehicle, anova analysis with Dunnett's post-test.

5-HT7 receptor activation increases PDGFβ receptor and PLCγ expression and reduces the expression of 5-HT7 receptors as shown in Fig. 1. Other 5-HT receptor subtypes have been shown to differentially effect the expression of NMDA receptor subunits. For example, in vivo, treatment with the 5-HT6 receptor ligand, SB-271047, differentially alters NMDA receptor subunit expression in the striatum (Marcos et al. 2010). To determine if the application of 5-HT7 receptor agonists altered NMDA receptor subunit expression in primary hippocampal neurons, we applied 300 nM LP 12 for 24 h. 5-HT7 receptor activation reduced the expression of the NR1 and NR2B subunits, but not the expression of NR2A (Fig. 4). The changes in NR1 and NR2B expression were blocked by the 5-HT7 receptor antagonist, SB 258719 (1 μM, Fig. 4a, b, e, f) and SB 269970 (1 μM, data not shown). Neither the neutral 5-HT7 receptor antagonist, SB 258719, nor the inverse agonist, SB 269970, changed NMDA receptor subunit expression on their own. To determine if the observed change in NR2B subunit expression was because of the activation of the PDGFβ receptor, we incubated the neurons with LP 12 in the absence or presence of STI-571. STI-571 prevented LP 12-induced decreases in both NR1 and NR2B subunit expression (Fig. 5).

Figure 4.

5-HT7 receptor activation decreases NR1 (a) and NR2B (e), but not NR2A (c), NMDA receptor subunit expression in primary hippocampal neurons. Primary hippocampal neurons were incubated with vehicle or 300 nM LP 12 (LP), 1 μM SB 258719 (SB 25), or both for 24 h. Western blots were quantified and bands were normalized to the loading control β-actin. Data are expressed as the fold change in the NR subunit expression versus vehicle. n = 9 (NR1), =12 (NR2A), = 14 (NR2B). Representative western blots are shown for each subunit in (b, d, and f). *< 0.05, anova analysis with Dunnett's post-test.

Figure 5.

Inhibition of platelet-derived growth factorβ receptor kinase activity blocks the LP 12-induced decrease in NR2B expression. (a, c) Primary hippocampal neurons were incubated with vehicle or 300 nM LP 12 (LP), 5 μM STI-571, or both for 24 h. Western blots were quantified and bands were normalized to the loading control β-actin. Data are expressed as the fold change in the NR1 (a, = 4) or NR2B (c, = 7)subunit expression versus vehicle, *< 0.05, anova analysis with Dunnett's post-test. Representative western blots are shown in (b and d).

Discussion

There is evidence that the brain itself uses the PDGF system as an endogenous neuroprotective mechanism to counter neuronal insults. Focal ischemia in rat brain causes an increase in PDGF-B mRNA that peaks at 24 h post ischemia (Iihara et al. 1994). PDGFβ receptor expression also rises rapidly after ischemia in rat brain (Iihara et al. 1996). In addition to PDGFβ receptor and PDGF ligand being up-regulated after ischemic events, exogenous PDGF-BB is also neuroprotective in focal ischemia (Sakata et al. 1998; Arimura et al. 2012) and against glutamate- and NMDA-induced cell death in cultured neurons (Egawa-Tsuzuki et al. 2004; Tseng and Dichter 2005; Beazely et al. 2009). Several studies have demonstrated that activation of synaptic pools of NMDA receptors promotes the phosphorylation of both cAMP-response element binding protein and extracellular signal-regulated kinase, and promotes cell survival, whereas the activation of extrasynaptic NMDA receptor pools promotes cell death (Soriano and Hardingham 2007). Others have suggested that it is the NMDA receptor composition, specifically the NR2 subunit, that dictates whether NMDA receptor activation will promote cell death (NR2B) or cell survival (NR2A) (Lai et al. 2011). We have previously demonstrated that PDGFβ receptors selectively inhibit NR2B-containing NMDA receptor currents, and that PDGFβ receptor neuroprotection against NMDA-induced toxicity is similar to that observed with the NR2B antagonist Ro 25-6981 (Beazely et al. 2009). In addition to the effects of PDGFβ receptors on NR2B-containing NMDA receptors, additional NMDA receptor-independent neuroprotective mechanisms downstream of PDGFβ receptors have been reported. These include the activation of the phosphoinositide 3-kinase (PI3K)/Akt pathway (Peng et al. 2008), increased expression of glutamate transporters on neurons and support cells (Figiel et al. 2003), and the involvement of transient receptor potential (TRP) C1 and TRPC6 channels (Yao et al. 2009a,b).

In this study, the kinase inhibitors STI-571 and AG1296 are both able to prevent LP 12-induced neuroprotection and changes in NMDA receptor subunit expression. The IC50 values for the inhibition of PDGFβ receptor kinase activity in vitro are in the 100-nM range; however, their IC50 values for inhibiting PDGFβ receptor-induced effects on, for example, cell growth, are in the low- to mid-μM range (Kovalenko et al. 1994; Roussidis et al. 2007; Weigel et al. 2009). We observed a significant effect on LP 12-induced neuroprotection at 1 and 5 μM for STI-571 and 5 μM for AG1296, and believe that this suggests that PDGFβ receptor kinase activity is required for the effects of LP 12. However, the interpretation of these results is made more difficult given the long-term (24 h) experimental design. LP 12 and the kinase inhibitors are coincubated over the entire 24-h period and to further complicate the interpretation, 5-HT7 receptors are being desensitized/down-regulated by 300 nM LP 12 (Fig. 1c) (Krobert et al. 2006) while PDGFβ receptor expression and phosphorylation state are increasing (Fig. 1a, b). Additional studies will be required to determine the temporal relationship between these events.

There is conflicting evidence for the effect(s) of 5-HT7 receptor on neuronal and synaptic activity. There is evidence that 5-HT7 receptors increase synaptic excitability in the hippocampus (Bacon and Beck 2000; Gill et al. 2002; Tokarski et al. 2005). However, other reports suggest that 5-HT7 receptors decrease glutamatergic signaling. Circadian phase advances in the dorsal raphe nucleus (DRN) involve an inhibition of glutamatergic neurotransmission (Duncan and Congleton 2010), and 5-HT7 receptors in the DRN antagonize NMDA receptor-dependent synaptic release of both serotonin and glutamate (Harsing et al. 2004). The glutamate-mediated increase in intracellular calcium is inhibited by 5-HT7 receptor activation in rat suprachiasmatic nucleus (Smith et al. 2001) and the amplitude of glutamate excitatory post-synaptic potentials (EPSPs) is decreased after 5-HT7 receptor activation in mouse suprachiasmatic nucleus (Quintero and McMahon 1999). Recently, Kobe et al. (2012) demonstrated that the activation of 5-HT7 receptors increased the number of dendrites and promoted synapse formation in primary hippocampal cultures, but did not affect the magnitude of long-term potentiation (LTP). However, 5-HT7 receptor knock-out mice display a reduced induction of LTP (Roberts et al. 2004). 5-HT7 regulation of PDGFβ receptor expression may explain the lack of clarity with respect to the effects of 5-HT7 receptors on glutamatergic signaling. In some of the studies cited above, 5-HT7 receptor agonists were applied for 5–10 minutes (Quintero and McMahon 1999; Harsing et al. 2004), whereas in others, animals/tissues were treated for hours to days, a time period sufficient for the up-regulation of PDGFβ receptors (Duncan and Congleton 2010; Kobe et al. 2012). Furthermore, none of the studies cited above used selective 5-HT7 agonists and only Kobe et al. (2012) used a selective 5-HT7 receptor antagonist in their study.

5-HT7 receptors display inverse agonist responses when treated with several antagonists (Romero et al. 2006). Furthermore, 5-HT7 receptors are rapidly desensitized in recombinant systems by both agonists and antagonists (Krobert et al. 2006), and indeed we saw a robust reduction in 5-HT7 receptor expression after prolonged exposure to the 5-HT7 receptor agonist, LP 12. Several antipsychotic agents with affinities for the 5-HT7 receptor, including respiridone, appear to “inactivate” 5-HT7 receptors (as measured by 5-HT-stimulated cyclic AMP accumulation in 5-HT7-receptor-expressing cells) after a brief exposure to these agents (Toohey et al. 2009). Toohey et al. (2009) hypothesize that these neuroleptics induce a stable 5-HT7 conformation that prevents receptor activation of Gαs. However, in the current study, neither the neutral antagonist, SB 258719, or the inverse agonist, SB 269970, changed PDGFβ receptor or NMDA receptor expression after being used alone. Similarly, in our previous work examining the effects of LP 12 on promoting PDGFβ receptor expression, neither of the 5-HT7 receptor antagonists used altered PDGFβ receptor expression when used alone (Vasefi et al. 2012).

Notwithstanding the effectiveness of growth factors to protect neurons against a variety of insults in vitro (Beazely et al. 2009) and in animal models (Wu et al. 2004), there are several obstacles that prevent their clinical use. For example, growth factors such as PDGF are several tens of kilodaltons in size and do not readily cross the blood–brain barrier. Certain growth factors including PDGF may increase the risk of glioma induction or proliferation (Shih and Holland 2006; Ellis et al. 2012a,b). Furthermore, PDGF causes several direct effects on the vasculature, including cerebral vasospasm in rats (Shiba et al. 2013). Could these obstacles be avoided by promoting the signaling of endogenous growth factors and their receptors in situ, rather than attempting to introduce exogenous proteins into the brain? Our results demonstrate that 5-HT7 receptor activation not only increases the expression of PDGFβ receptors but also increases the expression of PLCγ and causes a down-regulation of 5-HT7 receptors in primary hippocampal neurons. We demonstrate that long-term (24 h) activation of 5-HT7 receptors results in neuroprotective effects against NMDA-induced toxicity and that this effect is PDGFβ receptor dependent (Fig. 6). Furthermore, 5-HT7 agonists selectively decrease the expression of the NR2B and NR1 subunits of the NMDA receptor. The findings suggest that targeting G protein-coupled receptors may be a valid strategy to exploit the neuroprotective effects of growth factor receptors in neurons.

Figure 6.

Schematic displaying the proposed pathway whereby serotonin (5-HT)7 receptor agonists provide neuroprotection against NMDA toxicity via their effects on platelet-derived growth factor (PDGF)β receptor expression. Long-term (24 h) activation of 5-HT7 receptors increases the expression of PDGFβ receptors and PLCγ as well as an increase in the PLCγ-activating site on the PDGFβ receptor, Y1021 (green circle). 5-HT7 receptor activation prevents NMDA-induced toxicity and differentially regulates NMDA receptor subunit expression: NR1 and NR2B subunit expression is reduced, but NR2A subunit expression remains unchanged. Both the neuroprotective effects and the effects on NMDA receptor subunit expression induced by 5-HT7 agonists are blocked by PDGFβ receptor kinase inhibitors.

Acknowledgements

Special thanks to Nancy Gibson and Dawn McCutcheon for their animal care and use of their facility while ours was under construction. This research was funded through generous start-up funding from the University of Waterloo, Faculty of Science and by the National Science and Engineering Research Council of Canada. All the authors wish to declare that they have no conflict of interest.

Ancillary