Address correspondence and reprint requests to Alexandre Bouron, Laboratoire de Chimie et Biologie des Métaux, UMR CNRS 5249, CEA, 17 rue des Martyrs, 38054 Grenoble, France. E-mail: firstname.lastname@example.org
The lipid diacylglycerol (DAG) analogue 1-oleoyl-2-acetyl-sn-glycerol (OAG) was used to verify the existence of DAG-sensitive channels in cortical neurons dissociated from E13 mouse embryos. Calcium imaging experiments showed that OAG increased the cytosolic concentration of Ca2+ ([Ca2+]i) in nearly 35% of the KCl-responsive cells. These Ca2+ responses disappeared in a Ca2+-free medium supplemented with EGTA. Mn2+ quench experiments showed that OAG activated Ca2+-conducting channels that were also permeant to Ba2+. The OAG-induced Ca2+ responses were unaffected by nifedipine or omega-conotoxin GVIA (Sigma-Aldrich, Saint-Quentin Fallavier, France) but blocked by 1-[β-(3-(4-Methoxyphenyl)propoxy)-4-methoxyphenethyl]-1H-imidazole hydrochloride (SKF)-96365 and Gd3+. Replacing Na+ ions with N-methyl-d-glucamine diminished the amplitude of the OAG-induced Ca2+ responses showing that the Ca2+ entry was mediated via Na+-dependent and Na+-independent mechanisms. Experiments carried out with the fluorescent Na+ indicator CoroNa Green showed that OAG elevated [Na+]i. Like OAG, the DAG lipase inhibitor RHC80267 increased [Ca2+]i but not the protein kinase C activator phorbol 12-myristate 13-acetate. Moreover, the OAG-induced Ca2+ responses were not regulated by protein kinase C activation or inhibition but they were augmented by flufenamic acid which increases currents through C-type transient receptor potential protein family (TRPC) 6 channels. In addition, application of hyperforin, a specific activator of TRPC6 channels, elevated [Ca2+]i. Whole-cell patch-clamp recordings showed that hyperforin activated non-selective cation channels. They were blocked by SKF-96365 but potentiated by flufenamic acid. Altogether, our data show the presence of hyperforin- and OAG-sensitive Ca2+-permeable channels displaying TRPC6-like properties. This is the first report revealing the existence of second messenger-operated channels in cortical neurons.
C-type transient receptor potential protein family
The lipid diacylglycerol (DAG) is a second messenger involved in key cellular processes [for a review see (Carrasco and Merida 2007)]. Several DAG-sensitive proteins have been identified like some isoforms of the protein kinase C (PKC) family, Munc13 proteins (Brose et al. 2004) and ion channels of the C-type transient receptor potential protein family (TRPC). Seven TRPC are now identified and named TRPC1 to TRPC7. They all function as cation channels but they exhibit distinct gating and biophysical properties. For instance, homomeric TRPC1, TRPC3, TRPC6 or TRPC7 channels and heteromeric TRPC1–TRPC3 and TRPC3–TRPC4 channels can open in response to DAG application allowing an entry of Ca2+ into the cell (Hofmann et al. 1999; Lintschinger et al. 2000; Liu et al. 2005; Poteser et al. 2006). This DAG-dependent activation of TRPC channels occurs in a PKC-independent manner.
Several studies reported the existence of DAG-sensitive channels in neural cells. For instance, DAG activates Ca2+-conducting channels in the neuronal cell lines PC12 (Mwanjewe and Grover 2004) and IMR-32 (Nasman et al. 2006), as well as in cortical astrocytes (Grimaldi et al. 2003; Beskina et al. 2007), vomeronasal neurons (Lucas et al. 2003), hippocampal neurons (Tai et al. 2008) and neural stem cells (Pla et al. 2005). In astrocytes from embryonic rat brains, TRPC3 channels mediate the DAG-induced cytosolic Ca2+ changes (Grimaldi et al. 2003) whereas in astrocytes prepared from embryonic murine brains TRPC6 forms the DAG-sensitive channels (Beskina et al. 2007). On the other hand, the Ca2+ responses are due to TRPC2 in vomeronasal neurons (Lucas et al. 2003) and to TRPC6 in the hippocampus (Tai et al. 2008). These latter findings suggest that the molecular identity of the DAG-sensitive channels seems species- and tissue-dependent. At E13, the immature cortex expresses all TRPC isoforms (Boisseau, Kunert-Keil, Lucke and Bouron; unpublished data). We thus tried to determine whether the first cortical neurons, which are generated at E11–12 (Kriegstein and Noctor 2004), express functional DAG-sensitive channels. By recording cytosolic Ca2+ changes we observed that the DAG analogues 1-oleoyl-2-acetyl-sn-glycerol (OAG) or 1-stearoyl-2-arachidonoyl-sn-glycerol (SAG) caused a Ca2+ influx via channels sensitive to Gd3+ and SKF-96365 but insensitive to nifedipine and omega-conotoxin GVIA (ω-CTx), (Sigma-Aldrich). The OAG-induced Ca2+ responses were observed in KCl-responding and KCl-insensitive cells. Similarly to OAG or SAG, the DAG lipase inhibitor RHC80267, used to prevent the degradation of DAG, elevated [Ca2+]i but not the PKC activator phorbol 12-myristate 13-acetate (PMA). Moreover, the OAG-induced responses were not altered by PKC activation or inhibition. This shows that OAG recruited SKF-96365-sensitive Ca2+-conducting channels in a PKC-independent manner. Flufenamic acid (FFA) was used to further characterize the identity of the OAG-sensitive channels. FFA increases the amplitude of currents through TRPC6 channels but blocks TRPC3 and TRPC7 channels (Inoue et al. 2001). In cultured cortical neurons, FFA potentiated the OAG-induced Ca2+ responses suggesting that TRPC6 are key constituents of the OAG-sensitive channels. In situ hybridization experiments confirmed the presence of TRPC6, distributed throughout the cortical wall (preplate and ventricular zone) including in cortical neurons. Furthermore, Ca2+ imaging experiments and whole-cell patch-clamp recordings showed that hyperforin, a specific activator of TRPC6 channels (Leuner et al. 2007), activated non-selective cation channels blocked by SKF-96365. Currents through hyperforin-activated channels were increased by FFA. Altogether, these data suggest that OAG activated TRPC6 channels or channels exhibiting TRPC6-like properties. This is the first description of functional second messenger-operated channels in cortical neurons.
Materials and methods
Primary cell cultures
The cortical cells were dissociated from cerebral cortices isolated from embryonic (E13) C57BL6/J mice (vaginal plug was designated E0) according to procedures approved by the Ethical Committee of Rhône-Alpes Region (France). The cells have been prepared and kept up to 6 days in culture according to Bouron et al. (2006).
Calcium imaging experiments with Fluo-4
Cells were bathed in a Tyrode solution containing (in mM) 136 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 10 HEPES, 10 glucose, pH 7.4 (NaOH) and 1.8 μM Fluo-4/AM for 10 min at 20–22°C. They were then washed twice with a Fluo-4/AM-free Tyrode solution, stored 20 min at 20–22°C and then placed on the stage of an upright Olympus BX51WI microscope equipped with a water immersion 20× objective lens (Olympus, 0.95 NA, Sartrouville, France). The emitted light, provided by a 100 W mercury lamp, was attenuated by a neutral density filter (U-25ND6, Olympus). Fluorescent images were captured by a cooled digital CCD MicroMax Princetown camera (Roper Scientific, Evry, France). The software MetaFluor (Universal Imaging; Roper Scientific) was used to acquire the images at a frequency of 0.2 or 0.5 Hz and to analyse off-line the data. The shutter was controlled by the shutter driver Uniblitz VMM-D1 (Vincent Associates; Roper Scientific). The excitation light for Fluo-4 was filtered through a 460–495 nm excitation filter and the emitted light was collected through a 510–550 nm filter. The Ca2+ imaging experiments were performed 2–5 h after the plating of the cells and on cells kept up to 6 days in culture. When indicated, cells were maintained in a nominally Ca2+-free medium having the same composition as the Tyrode solution (see above) supplemented with 0.4 mM EGTA. The pH of the EGTA-containing solution was adjusted to pH 7.4 with NaOH. In some experiments, cells were stimulated with a K+-rich solution containing 50 mM (instead of 5 mM) KCl. Under these conditions, the concentration of NaCl was reduced to 91 mM. All solutions were applied through a gravity-driven system perfusing the entire recording chamber. The fluorescence was collected from 10–40 cells simultaneously monitored. Only one field of view was used per dish. Data are presented as mean ± SEM, with n being the number of cell bodies tested. The experiments reported below were carried out at 20–22°C.
Calcium imaging experiments with fura-2
The uneven distribution of Fluo-4 and its photo-bleaching can limit the use of this Ca2+ indicator. To avoid these problems we used the ratiometric dye fura-2. Cells grown on 15 mm diameter glass cover-slips were incubated in a Tyrode solution supplemented with 2.5 μM fura-2 for 15 min at 20–22°C. They were then washed twice and kept in a fura-2-free Tyrode solution for 20 min at 20–22°C. Cover-slips were transferred on a perfusion chamber (RC-25F, Warner Instruments; Phymep, Paris, France) and placed on the stage of an Axio Observer A1 microscope (Carl Zeiss, Sartrouville, France) equipped with a CoolSnap HQ2 camera (Princeton Instruments; Roper Scientific) and a Fluar 40× oil immersion objective lens (1.3 NA) (Carl Zeiss). Light was provided by the DG-4 wavelength switcher (Princeton Instruments). A dual excitation at 340 and 380 nm was used and emission was collected at 515 nm. The software MetaFluor (Universal Imaging) was used to acquire the images at a frequency 0.5 Hz and to analyse off-line the data. Stock solutions of OAG, SAG, nifedipine, hyperforin, cytochalasin D, PP2, RHC80267, GF 109203X and genistein were prepared in dimethyl sulfoxide (DMSO) and diluted at least 1000-fold into the Tyrode solution immediately before use so that the final concentration of DMSO never exceeded 0.1%. Control experiments were performed with DMSO alone (0.1%). At this concentration, the solvent never induced any cytosolic Ca2+ signal (not shown). All other stock solutions (SKF-96365, ω-CTx, Gd3+, bethanechol, phenylephrine, histamine hydrochloride and α-methyl-5-hydroxytryptamine) were prepared in water and also diluted at least 1000-fold into the Tyrode solution immediately before use.
Sodium imaging experiments with CoroNa Green
The effect of OAG on the cytosolic concentration of Na+ was assayed with the fluorescent Na+ indicator CoroNa Green. Cells were incubated with 5 μM CoroNa Green/AM for 30 min at 20–22°C. They were rinsed and placed on the stage of an Axio Observer A1 microscope (Carl Zeiss). These experiments were carried out with the fura-2 Ca2+ imaging setup described above, except that a single excitation at 495 nm was used and emission was collected at 525 nm.
The whole-cell configuration of the patch-clamp technique (Hamill et al. 1981) was used to record currents activated by hyperforin. The experiments were conducted according to protocols and recording solutions already described (Hill et al. 2006). The external medium contained (in mM): 140 NaCl, 4 KCl, 10 TEACl, 1 CaCl2, 1 MgCl2, 10 HEPES, 5 D-glucose, pH 7.4 (NaOH). The patch pipettes were fabricated by means of the DMZ Universal pipette puller (Zeitz Instruments, München, Germany) from thick wall borosilicate glass capillaries (1.5 mm o.d. × 0.86 mm i.d., Clark Electromedical Instruments; Phymep). When filled with the following intracellular solution (in mM): 140 CsCl, 1 MgCl2, 5 d-glucose, 10 HEPES, 0.1 EGTA, 4 ATP, 0.2 GTP, pH 7.2 (CsOH), pipettes had a resistance of 2.5–3.8 MΩ. Currents were measured with an Axoclamp 200B amplifier (Axon Instruments, Dipsi, Chatillon, France), filtered at 1 kHz, and analysed off-line using the pClamp software (version 9.0, Axon Instruments). Unless otherwise indicated, the holding membrane potential was set at 0 mV to inactivate voltage-gated Na+ and Ca2+ channels. Whole-cell currents, recorded at 20–22°C 1–3 days after the plating of the cells, were triggered at a frequency of 0.2 Hz by 2 s voltage ramps from −100 to +40 mV. Capacitive transients were cancelled and the cell capacitance value was read from the amplifier dials.
Total RNA was isolated using guanidinium-isothiocyanate (RNeasy Mini Kit; Qiagen, Hilden, Germany) and RNA concentration was determined by UV absorbance measurements. An amount of 200 ng total RNA was reverse transcribed using random hexamer primers and the TaqMan Reverse Transcription Reagents (PE Applied Biosystems, Weiterstadt, Germany). PCR was performed with Taq Polymerase for 40 cycles as described previously (Kunert-Keil et al. 2006). Amplification of a TRPC6 cDNA fragment was done with murine specific TRPC6 primers (Assay-on-Demand Mm00443441_m1; PE Applied Biosystems).
In situ hybridization
Non-radioactive in situ hybridization was performed with cryo sections (4 μm) which had been fixed in 4% paraformaldehyde. Sections were rehydrated and permeabilized with 0.2 M HCl. Post-fixation (paraformaldehyde 4%, 20 min, 4°C) was followed by acetylation using 0.4% acetic anhydride in triethanolamine (0.1 M, pH 8.0, 15 min). After washing with 50% formamide in 1.5% sodium-sodium phosphate-EDTA buffer (20x sodium-sodium phosphate-EDTA buffer: 3.6 M NaCl, 0.2 M NaH2PO4, 0.2 M EDTA, pH 7.4) the sections were pre-hybridized for 1 h at 56°C in a solution containing 50% formamide and 50% solution D (4 M guanidine thiocyanate, 25 mM sodium citrate, pH 7.0), 0.5% blocking reagent (Roche Biochemicals, Mannheim, Germany) and 210 μg/mL t-RNA. For TRPC6, a cDNA fragment (GenBank: NM_013838; position nt +2181 to + 2507) was cloned into the pGEM-T-Easy cloning vector (Promega, Mannheim, Germany) (Kunert-Keil et al. 2006). After hybridization for 12–16 h with pre-hybridization solution containing 125 ng digoxigenin (DIG)-labelled cRNA probe (Kunert-Keil et al. 2006) and washing with 2x saline sodium citrate buffer (20x saline sodium citrate buffer: 3 M NaCl, 0.3 M sodium citrate; pH 7.4) sections were incubated with blocking reagent (Roche Biochemicals). Bound riboprobe was visualized by incubation with alkaline phosphatase-conjugated anti-DIG antibody (Roche Biochemicals) and subsequent substrate reaction containing 5-bromo-4-chloro-3-indolyl phosphate/nitroblue-tetrazolium chloride.
GF 109203X (bisindolylmaleimide I (2-[1-[3-(dimethylamino)propyl]-1H-indol-3-yl]-4-(1H-indol-3-yl)-1H-pyrrole-2,5-dione, or Gö 6850) and PP2 were purchased from Calbiochem (VWR, Fontenay sous Bois, France). The fluorescent Ca2+ and Na+ indicators Fluo-4, fura-2, CoroNa Green were from Molecular Probes (Interchim, France). The alkaline phosphatase-conjugated anti-DIG antibody and the blocking reagent were from Roche Biochemicals. All other chemicals including PMA, OAG, SAG, FFA, RHC80267 and N-methyl-d-glucamine (NMDG), genistein, methyl-β-cyclodextrin (MβCD), cytochalasin D, bethanechol, phenylephrine, serotonin, α-methyl-5-hydroxytryptamine were from Sigma-Aldrich. Hyperforin is a kind gift from Dr Willmar Schwabe GmbH & Co (Karlsruhe, Germany).
OAG evoked Ca2+ responses
Most of the freshly dissociated or cultured E13 cortical cells express βIII-tubulin, a marker of early post-mitotic neurons (Bouron et al. 2005) and possess functional voltage-gated Ca2+ channels (Bouron et al. 2006). The external application of OAG (100 μM) evoked Ca2+ responses in some KCl-responding and KCl-insensitive cells as shown in Fig. 1. In this report, we only analysed the properties of cells where a KCl-induced depolarization evoked a transient Ca2+ rise. These KCl-responsive cells were considered as neurons (Bouron et al. 2006). In contrast to the KCl responses, the OAG-induced Ca2+ responses were asynchronous and displayed a diversity of shapes as illustrated in Fig. 2(a). Similar experiments were carried out after 1, 2, 4, 5 and 6 days in vitro (DIV). OAG evoked Ca2+ responses in freshly dissociated cells (Fig. 2a) as well as in cells kept up to 6 days in primary culture (data not shown). These Ca2+ responses had roughly the same amplitude regardless of the age of the cells, at least up to six DIV. The total number of KCl-responsive cells used in the present study was 1129. Only 389 of them (35%) displayed OAG-induced Ca2+ responses. Schematically, based on their time course three types of OAG-induced Ca2+ responses were identified: (i) cells with transient Fluo-4 responses (15%), (ii) cells exhibiting a sustained elevation of Ca2+ (23%) and (iii) cells with oscillatory Ca2+ responses (62%, total 100%). These three phenotypes could be seen in distinct cells from the same dish, whether freshly dissociated cells or cells kept several days in culture.
OAG controlled an entry of Ca2+ and Na+
The OAG-induced Ca2+ responses were not observed when the cells were bathed in a Ca2+-free medium (supplemented with 0.4 mM EGTA) (n = 65 cells) (Fig. 2b). Thus, OAG did not release Ca2+ from stores but rather promoted an entry of Ca2+. Furthermore, the OAG-induced Ca2+ responses were unaffected by the voltage-gated Ca2+ channel inhibitors nifedipine (5 μM, n = 40) or ω-CTx (1 μM, n = 37) but they were strongly attenuated in the presence of Gd3+ (5 μM, n = 106) and SKF-96365 (20 μM, n = 82) (Fig. 2b). A recent study showed that DAG can promote an entry of Na+ through TRPC6 channels which, in turn, elevates the cytosolic concentration of Ca2+ via the Na+/Ca2+ exchanger operating in the reverse mode (Lemos et al. 2007). As Gd3+ blocks the Na+/Ca2+ exchanger (Zhang and Hancox 2000), experiments were performed with a Na+-free Tyrode solution where Na+ was replaced with the large organic cation NMDG+. Under these conditions, OAG-induced Ca2+ responses were still observed but, on average, their amplitudes were reduced by nearly 50% (n = 192) (Fig. 2b). This latter result thus suggests that two mechanisms contributed to the OAG-induced Ca2+ entry: OAG triggers a Gd3+- and SKF-96365-sensitive Ca2+ rise via Na+-dependent and Na+-independent mechanisms. The effects of OAG on [Na+]i was assayed by means of the fluorescent Na+ indicator CoroNa Green (Poburko et al. 2007). Figure 2(c) shows representative CoroNa Green responses obtained in three different experimental conditions. OAG induced a time-dependent increase in CoroNa Green fluorescence, an effect that was not observed in the presence of Gd3+ or when Na+ was replaced with NMDG+ (Fig. 2c and d). Thus, OAG elevated [Na+]i via a Gd3+-sensitive process. This response depends on the extracellular concentration of Na+.
Some experiments were performed with the ratiometric Ca2+ indicator fura-2 and with SAG, another DAG analogue. SAG induced Ca2+ transients (Fig. 2e). Similar fura-2 signals were obtained with OAG (not shown). When replacing 2 mM CaCl2 with 2 mM BaCl2, SAG elevated the cytosolic concentration of Ba2+ (Fig. 2e). These Ba2+ signals were sustained, not transient as seen with Ca2+ because Ba2+ is weakly pumped out of the cells or into the organelles (Yamaguchi et al. 1989). To further verify the presence of DAG-activated Ca2+ channels, we checked whether Mn2+ could quench the fura-2 fluorescence. Mn2+ ions readily enter cells through Ca2+-conducting channels. The results of these experiments are summarized in Fig. 2(f). Cells were kept in a Ca2+-free medium. The addition of Mn2+ (100 μM) weakly diminished the fura-2 fluorescence at 340 and 380 nm. However, a subsequent addition of SAG was accompanied by a clear quench of the fura-2 fluorescence at both wavelengths (Fig. 2f). Altogether, these experiments showed that DAG analogues like OAG or SAG activated SKF-96365-sensitive channels permeable to Ca2+, Ba2+ and Mn2+. Based on our data, it is proposed that in embryonic cortical neurons OAG activates SKF-96365-sensitive cation channels allowing an entry of Ca2+ and Na+. This OAG-induced Na+ entry thus stimulates the Na+/Ca2+ exchanger which, when operating in the reverse mode, allows an entry of Ca2+. According to this hypothesis, OAG elevates the cytosolic concentrations of Ca2+ by activating OAG-sensitive Ca2+-conducting channels and also by promoting an entry of Ca2+ via the Na+/Ca2+ exchanger. These results are in agreement with recent data obtained on smooth muscle cells (Poburko et al. 2007; Syyong et al. 2007). It is important to note that the Na+/Ca2+ exchanger is already functional at E13 and mediates a Na+-dependent Ca2+ entry (Platel et al. 2005). Both responses, namely the Na+ and Ca2+ entries, disappeared in the presence of Gd3+, a potent blocker of calcium channels and of the Na+/Ca2+ exchanger (Zhang and Hancox 2000).
The DAG lipase inhibitor RHC80267 induced Ca2+ responses
The DAG lipase inhibitor RHC80267 (50 μM), used to prevent the degradation of DAG, can elevate [Ca2+]i (Hofmann et al. 1999). When acutely applied, RHC80267 elicited a weak Fluo-4 increase (8%, n = 133 cells) (Fig. 3a). However, this RHC80267 treatment, done at 20–22°C, was probably too short to significantly alter the intracellular concentration of DAG and therefore to promote robust Ca2+ responses. Experiments were then realized after a longer RHC80267 treatment during which the cells were maintained at 37°C. This was performed as follows: Fluo-4-loaded cells were kept in a Ca2+-free medium containing 0.4 mM EGTA + 50 μM RHC80267 for 20 min at 37°C. They were then placed on the stage of the microscope and superfused with a recording medium containing 2 mM Ca2+. The readmission of Ca2+ was accompanied by large Fluo-4 signals (Fig. 3b). When similar experiments were conducted with RHC80267-untreated cells, the Ca2+ challenge was followed by much smaller Fluo-4 signals (Fig. 3b) which most likely reflect an entry of Ca2+ through store-operated channels. Thus, an entry of Ca2+ can be triggered by either inhibiting the DAG lipase with RHC80267 or by applying OAG (or SAG).
Protein kinase C neither mimicked nor affected OAG responses
Many cellular effects of DAG have been attributed to PKC, its major downstream effector. Among other responses, PKC can lead to cytosolic Ca2+ changes (Khoyi et al. 1999; Murthy et al. 2000; Rosado and Sage 2000; Albert and Large 2002). In order to determine whether OAG acted via a PKC-dependent mechanism, the phorbol ester PMA, a potent PKC activator, was used. The external application of PMA (1 μM) did not elevate [Ca2+]i (Fig. 3c). We next tried to verify whether PKC regulated the OAG-induced Ca2+ entry. Figure 3(d) shows that PKC activation (with PMA), or PKC inhibition (with the PKC antagonist GF 109203X, also named bisindolylmaleimide I or Gö 6850) did not alter the OAG-induced Ca2+ entry. Thus, OAG promoted a PKC-independent entry of Ca2+.
Flufenamic acid potentiated the OAG responses
In neural cells, the presence of a PKC-independent but OAG-sensitive entry of Ca2+ has been described in astrocytes (Grimaldi et al. 2003; Beskina et al. 2007), vomeronasal neurons (Lucas et al. 2003), hippocampal neurons (Tai et al. 2008) and neural stem cells (Pla et al. 2005). Depending on the tissue and the species (rat vs. mouse), TRPC2 (Lucas et al. 2003), TRPC3 (Grimaldi et al. 2003) or TRPC6 (Beskina et al. 2007; Tai et al. 2008) are the key elements of these OAG-sensitive channels but whatever their exact molecular identity, TRPC appear as likely candidates controlling OAG-sensitive channels (Hofmann et al. 1999; Lintschinger et al. 2000; Liu et al. 2005). FFA is an anti-inflammatory agent which, among other cellular actions, inhibits currents through TRPC3 and TRPC7 channels but increases currents through TRPC6 channels (Inoue et al. 2001; Jung et al. 2002). We thus took advantage of this property to further characterize the OAG-sensitive channels. Applied alone (without OAG), FFA produced a modest but long-lasting elevation of the cytosolic concentration of Ca2+. The Fluo-4 fluorescence increased by nearly 25% in the presence of FFA (n = 14 cells, p < 0.05, Student’s t-test) (Fig. 4). When compared with OAG, FFA gave rise to weak Ca2+ responses having distinct kinetics properties. The FFA-induced Ca2+ rise also occurred when the cells were bathed in a Ca2+-free medium (not shown) indicating that FFA promoted the release of Ca2+ from internal stores as already shown in non-neuronal (Poronnik et al. 1992; Cruickshank et al. 2003) and in neuronal cells (Lee et al. 1996). When added in the presence of FFA, OAG triggered larger Fluo-4 responses when compared with OAG alone (Fig. 4a). The amplitude of the Fluo-4 signals seen in the presence of FFA + OAG was larger than the sum of the two signals (the FFA-induced Ca2+ release + the OAG-induced Ca2+ entry). In addition, with FFA, the number of OAG responsive cells was > 60% instead of ∼30% without FFA. Instead of blocking the OAG-induced Ca2+ entry, the channel blocker FFA increased the number of OAG responsive cells and potentiated the OAG-induced Ca2+ entry. As TRPC6 is the only known TRPC channel of which activity is up-regulated by FFA (Inoue et al. 2001; Jung et al. 2002), it seems that OAG activates TRPC6 channels or channels exhibiting TRPC6-like properties. We then addressed the question of the presence of TRPC6 in the cortex at E13.
TRPC6 was found in the cortex of E13 mice
Using standard PCR, TRPC6 transcripts of expected size (153 bp) were detected in the brain and cortex from embryonic (E13) mice as shown in Fig. 4(c). The PCR products were exemplary sequenced and revealed the expected DNA sequence. In situ hybridization experiments with the antisense TRPC6 cRNA probe were carried out to better describe the expression of TRPC6 mRNA. It was found throughout the cortex, both in the preplate and in the ventricular zone (Fig. 4c). Sense probes applied as controls did not show positive results.
Hyperforin triggered an entry of Ca2+
Cortical neurons express TRPC6 channels and OAG activates a Ca2+ entry pathway exhibiting TRPC6-like properties (up-regulation by FFA). If TRPC6 channels mediate these OAG-sensitive Ca2+ responses, hyperforin should mimic the effect of OAG. Indeed, hyperforin, the main active principle of St John’s wort extract, specifically activates TRPC6 channels without activating TRPC1, TRPC3, TRPC4 and TRPC5 channels (Leuner et al. 2007). Like OAG, hyperforin triggers a massive entry of Ca2+ blocked by Gd3+ (Fig. 4d and e).
Hyperforin activated a non-selective cation current
When held at a holding potential of −50 mV, the external application of hyperforin elicited an inward current that transiently increased and then declined to baseline (Fig. 5a). On average, when measured at −50 mV, the maximal amplitude of the current induced by 5 μM hyperforin was 125 ± 17 pA (n = 14). In HEK293 cells over-expressing TRPC6 channels (Leuner et al. 2007) like in cortical neurons, hyperforin triggers a transient inward current. In neurons, this current was reduced by ∼80% in the presence of 20 μM SKF-96365 (n = 4, p < 0.01) (Fig. 5a) or 10 μM Gd3+ (n = 5) (not shown). Figure 5(b) shows representative current–voltage relationships obtained before and after the addition of 5 μM hyperforin. When recorded with a caesium-rich pipette solution, this antidepressant elicited an inward and outward current having a reversal potential near 0 mV, indicating that hyperforin recruited non-selective cation channels. In addition, when added after hyperforin, FFA potentiated the current (n = 5/5 cells tested) (Fig. 5c). This enhancement was specific as its vehicle never affected the hyperforin-activated current (n = 5 cells tested). Thus, hyperforin activated SKF-96365- and Gd3+-sensitive cation channels that were positively regulated by FFA.
Properties of the OAG-sensitive channels
The characterization of these cation channels was further investigated by comparing their properties to those of TRPC6 channels (Dietrich and Gudermann 2007). The activity of TRPC6 channels is under the control of PP2-sensitive src protein tyrosine kinases (Hisatsune et al. 2004; Aires et al. 2006). For instance, treating HEK293 cells over-expressing TRPC6 channels with the tyrosine kinase inhibitor PP2 attenuates the DAG-activated Ca2+ entry (Aires et al. 2006). It is however worth adding that TRPC6 channels over-expressed in HEK293 cells are insensitive to genistein, another tyrosine kinase inhibitor (Kawasaki et al. 2006). Cortical cells were incubated with PP2 (n = 29 cells) or genistein (n = 43) before adding OAG. Neither of these tyrosine kinase inhibitors affected the OAG-induced Ca2+ responses (Fig. 5d). Several studies have highlighted the importance of caveolae in Ca2+ homeostasis (Isshiki and Anderson 2003). The entry of Ca2+ through TRPC6 channels over-expressed in HEK293 cells is regulated by an exocytotic mechanism with TRPC6 channels present in caveolae-related microdomains (Cayouette et al. 2004). The application of OAG promotes the insertion of TRPC6 channels into the plasma membrane (Cayouette et al. 2004). Treating HEK293 cells over-expressing TRPC6 channels with MβCD, which disrupts lipid rafts, causes a complete suppression of the DAG-activated Ca2+ entry (Aires et al. 2006). Similar experiments were conducted with cortical neurons. MβCD did not affect the OAG-induced Ca2+ responses (n = 98 cells) (Fig. 5d). In another series of experiments, we verified whether the OAG-induced Ca2+ responses were controlled by an actin-dependent trafficking step. Cortical neurons were treated with cytochalasin D, a membrane-permeant inhibitor of actin polymerization. Here again, the OAG-induced Ca2+ responses were unaffected by cytochalasin D (n = 41 cells) (Fig. 5d). In a final set of experiments, cells were stimulated with one of the following neurotransmitter receptor agonists to gain further insight into the physiological relevance of these OAG-induced Ca2+ responses: bethanechol (10–100 μM, n = 50 cells), α-methyl-5-hydroxytryptamine (10 μM, n = 47 cells), phenylephrine (10 μM, n = 63 cells) and histamine (10 μM, n = 52 cells). These agonists recruit, respectively, muscarinic acetylcholine, serotonin 5-HT2, α1-adrenergic, and histamine receptors and promote the production of DAG. None of these agonists tested induced Ca2+ responses (not shown). Although the identity of the signalling pathway controlling the DAG-activated cationic channels remains unclear, this report clearly shows the existence of functional second messenger-operated cationic channels in cortical neurons from E13 mouse embryos.
By performing Ca2+ imaging and electrophysiological experiments we have shown that in cortical cells from E13 mouse embryos OAG (or SAG) and hyperforin activate plasma membrane cation channels. These OAG-induced Ca2+ responses were observed in KCl-responsive and in KCl-unresponsive cells. This latter cell population, considered as non-neuronal cells, was not further analysed as recent studies already demonstrated that OAG triggers an entry of Ca2+ in cortical astrocytes. It develops via TRPC3 channels in cortical astrocytes prepared from E17 rat embryos (Grimaldi et al. 2003) and via TRPC6 channels in cortical astrocytes prepared from E17 mouse embryos (Beskina et al. 2007). In addition, an OAG-induced Ca2+ entry is found in neural stem cells prepared from E13 rat embryos (Pla et al. 2005). Thus, our data, showing a OAG-induced Ca2+ entry in KCl-unresponsive cells (e.g. cell 2 in Fig. 1) is in line with reports describing OAG-induced Ca2+ signals in non-neuronal cortical cells (Grimaldi et al. 2003; Beskina et al. 2007). Therefore, we focused our analysis on the KCl-responsive cells. To our knowledge, this is the first report showing second messenger-operated channels in cortical neurons.
Nearly 35% of the KCl-responsive cells responded to OAG. In most cases, OAG induced Ca2+ oscillations even in cells treated with caffeine to deplete the caffeine-sensitive Ca2+ pool of the endoplasmic reticulum (not shown). The OAG-induced Ca2+ signals, seen in freshly dissociated cells as well as in cultured isolated cells kept up to six DIV, disappeared when Ca2+ was omitted from the extracellular milieu. The entry of Ca2+ could be triggered by OAG, SAG or in the presence of the DAG lipase inhibitor RHC80267. This Ca2+ route was unaffected by the voltage-gated Ca2+ channel antagonists nifedipine and ω-CTx but it was strongly blocked by Gd3+ and SKF-96365. Replacing Na+ ions with NMDG+ did not suppress but attenuated the OAG-induced Ca2+ rise. Analysis of cytosolic Na+ changes with the fluorescent Na+ indicator CoroNa Green revealed that OAG caused an entry of Na+. A similar observation was made in vascular smooth muscle cells (Poburko et al. 2007). Based on these findings it is proposed that DAG controls the activity of SKF-96365-sensitive channels allowing a Ca2+ entry via Na+-dependent and Na+-independent mechanisms. As the cytosolic Ca2+ rise partially depends on the extracellular concentration of Na+, OAG controls the activity of Na+- and Ca2+-conducting channels. The intracellular load of Na+ thus activates the Na+/Ca2+ exchanger which, in turn, permits an entry of Ca2+ (Platel et al. 2005). This DAG-induced Ca2+ entry was blocked by Gd3+. However, the broad spectrum Ca2+ channel antagonist Gd3+ also inhibits the Na+/Ca2+ exchanger (Zhang and Hancox 2000) and the OAG-induced elevation of Na+. PKC, a major downstream DAG effector, can increase the cytosolic concentration of Ca2+ in some cell types (Albert et al. 1987; Khoyi et al. 1999; Murthy et al. 2000; Rosado and Sage 2000; Albert and Large 2002). In contrast to OAG or SAG, the PKC activator PMA failed to trigger any Ca2+ response. Furthermore, stimulating or inhibiting PKC activity had no effect on the DAG-induced Ca2+ entry. Taken together, these results favour the existence of cation channels activated by DAG in a PKC-independent manner.
TRPC are Gd3+- or SKF-96365-sensitive plasma membrane proteins forming voltage-independent cation channels. Some isoforms constitute DAG-sensitive Ca2+-conducting channels. When heterogeneously expressed, homomeric TRPC3, TRPC6 or TRPC7 channels and heteromeric TRPC1–TRPC3, TRPC3–TRPC4 channels function as DAG-sensitive Ca2+-conducting channels (Hofmann et al. 1999; Lintschinger et al. 2000; Liu et al. 2005; Poteser et al. 2006). Such DAG-sensitive channels have been described in the neuronal cell lines PC12 and IMR-32 (Mwanjewe and Grover 2004; Nasman et al. 2006) as well as in neural cells like cortical astrocytes (Grimaldi et al. 2003; Beskina et al. 2007), vomeronasal neurons (Lucas et al. 2003), hippocampal neurons (Tai et al. 2008) and neural stem cells (Pla et al. 2005). TRPC3 (in rats) (Grimaldi et al. 2003) or TRPC6 channels (in mice) (Beskina et al. 2007) participate in the DAG-induced cytosolic Ca2+ changes in astrocytes but the Ca2+ entry occurs through TRPC2 channels in vomeronasal neurons (Lucas et al. 2003) and through TRPC6 channels in hippocampal neurons (Tai et al. 2008). Based on these findings, we suggest that TRPC channels or channels exhibiting TRPC-like properties are involved in the DAG-dependent Ca2+ entry of cortical neurons. Among the various TRPC isoforms described so far, only currents through TRPC6 channels are increased by FFA. This anti-inflammatory drug is a potent blocker of anion and cation channels including TRPC (Inoue et al. 2001). But heterogeneously expressed TRPC6 channels (in HEK cells) and native TRPC6 channels (in vascular smooth muscle cells) are up-regulated by FFA (Inoue et al. 2001; Jung et al. 2002). In smooth muscle cells, FFA inhibits native Ca2+-conducting channels having TRPC3- and TRPC7-like properties with an IC50 value of 2.45 μM (Peppiatt-Wildman et al. 2007) but increases currents through native TRPC6-like channels (Hill et al. 2006). On the other hand, based on their electrophysiological and pharmacological studies, Carter et al. (2006) concluded that TRPC6 was involved in the ADP-dependent cation influx of murine megakaryocytes (Carter et al. 2006). Interestingly, this response was strongly enhanced by FFA (Carter et al. 2006). Therefore, TRPC6 (Inoue et al. 2001; Jung et al. 2002) and TRPC6-like channels (Carter et al. 2006; Hill et al. 2006) appear as the only TRPC channels up-regulated by FFA. In E13 cortical neurons, the OAG-induced Ca2+ entry was enhanced by FFA. Furthermore, hyperforin which selectively activates TRPC6 channels without activating TRPC1, TRPC3, TRPC4 and TRPC5 (Leuner et al. 2007) mimics the action of OAG. Electrophysiological measurements showed that this antidepressant activated non-selective cation channels blocked by Gd3+ (not shown) and SKF-96365. This in line with a recent study showing that hyperforin selectively activates non-selective cation TRPC6 channels (Leuner et al. 2007). In addition, hyperforin-activated currents were increased by FFA.
On the other hand, if the DAG-sensitive channels of cortical neurons exhibit TRPC6-like characteristics, they however display properties that are not found in other cells expressing TRPC6 channels. For instance, in HEK cells over-expressing TRPC6 channels, disruption of lipids rafts abolishes Ca2+ entry through TRPC6 channels (Aires et al. 2006) whereas the same treatment has no effect in cortical neurons. It is however possible that native and over-expressed TRPC6 channels exhibit distinct properties as already shown for TRPC3 where the mode of regulation of this TRPC isoform critically depends on its level of expression (Putney 2004). Another property of the DAG-sensitive channels of cortical neurons is their insensitivity to the tyrosine kinase inhibitors PP2 and genistein. Indeed, the src tyrosine kinase inhibitor PP2 abolishes endogenous TRPC6-dependent Ca2+ entry in cardiac myocytes (Nishida et al. 2007) and in HEK cells over-expressing TRPC6 channels (Aires et al. 2006) but has no effect on cortical neurons. Of note, Kawasaki et al. (2006) also reported that TRPC6 channels were unaffected by the tyrosine kinase inhibitor genistein.
Although the molecular identity of the DAG-sensitive channels of cortical neurons is not firmly established we exclude TRPC3 as the main candidate. This is based on the experiments carried out with PMA and showing that, after a PMA treatment, OAG was still able to promote an entry of Ca2+. Indeed, PKC activation totally blocks TRPC3 in response to OAG (Trebak et al. 2003; Venkatachalam et al. 2003; Kwan et al. 2006). A key issue concerns the characterization of the physiological activator(s) of these channels. None of the neurotransmitter receptor agonist tested (bethanechol, α-methyl-5-hydroxytryptamine, phenylephrine and histamine) had an effect. Although a clear understanding of the signalling pathway controlling the DAG-activated cation channels remains unclear as well as their exact molecular identity, we provide experimental evidence for the existence of functional second messenger-operated cationic channels in cortical neurons from E13 mouse embryos. Inositol 1,4,5-trisphosphate and DAG are second messengers playing important roles in cell signalling. Inositol 1,4,5-trisphosphate links cell surface receptors and Ca2+ signalling whereas DAG is the physiological activator of protein kinase C and thus controls protein phosphorylation. This latter process is regarded as one of the most important molecular mechanisms by which extracellular signals produce their biological responses (Walaas and Greengard 1991). As already shown, DAG also regulates in a PKC independent manner the activity of some plasma membrane ion channels (Hofmann et al. 1999; Lintschinger et al. 2000; Poteser et al. 2006). By controlling the activity of various ion channels and the phosphorylation of a plethora of proteins, DAG is as a second messenger with a widespread biological importance.
We wish to thank D. Poburko for helpful comments and suggestions on a previous version of this work. We also wish to thank Dr Willmar Schwabe (Karlsruhe, Germany) for the kind gift of hyperforin. This study was supported by a grant from l’Agence Nationale de la Recherche (ANR-2006-SEST).