|Home | About | Journals | Submit | Contact Us | Français|
The glycerol-based lysophospholipid lysophosphatidylinositol (LPI) is an endogenous agonist of the G-protein-coupled receptor 55 (GPR55) exhibiting cannabinoid receptor-like properties in endothelial cells. To estimate the contribution of GPR55 to the physiological effects of LPI, the GPR55-dependent and -independent electrical responses in this cell type were investigated.
Applying small interference RNA-mediated knock-down and transient overexpression, GPR55-dependent and -independent effects of LPI on cytosolic free Ca2+ concentration, membrane potential and transmembrane ion currents were studied in EA.hy296 cells.
In a GPR55-dependent, GDPβS and U73122-sensitive manner, LPI induced rapid and transient intracellular Ca2+ release that was associated with activation of charybdotoxin–sensitive, large conductance, Ca2+-activated, K+ channels (BKCa) and temporary membrane hyperpolarization. Following these initial electrical reactions, LPI elicited GPR55-independent long-lasting Na+ loading and a non-selective inward current causing sustained membrane depolarization that depended on extracellular Ca2+ and Na+ and was partially inhibited by Ni2+ and La3+. This inward current was due to the activation of a voltage-independent non-selective cation current. The Ni2+ and La3+-insensitive depolarization with LPI was prevented by inhibition of the Na/K-ATPase by ouabain.
LPI elicited a biphasic response in endothelial cells of which the immediate Ca2+ signalling depends on GPR55 while the subsequent depolarization is due to Na+ loading via non-selective cation channels and an inhibition of the Na/K-ATPase. Thus, LPI is a potent signalling molecule that affects endothelial functions by modulating several cellular electrical responses that are only partially linked to GPR55.
Lysophospholipids (LPs) have been recognized as potent signalling molecules that modulate many cell functions in a variety of tissues including the cardiovascular, immune, nervous, and reproductive systems (Gardell et al., 2006). The generation of these signalling molecules is most often linked to the metabolism of membrane phospholipids by enzymes that are located at the inner side of the plasma membrane and are activated upon cell stimulation. Initially, LPs have thus been discussed to serve as second-messenger molecules that modulate intracellular signalling events (Corda et al., 2002). However, recent findings point to LPs also as intercellular signalling molecules that are involved in cell-to-cell communication. This assumption is based on reports that LPs are transported to, and even produced in, the extracellular space and that cell surface receptors that are specifically activated by certain LPs have been identified.
However, the diverse effects of different LPs as signalling molecules are only marginally explored on the molecular level, although they deserve particular interest in order to understand the molecular processes of lipotoxic pathways. Increased levels of LPs have been found particularly under pathological conditions in the vasculature, such as atherosclerosis (Ridgway et al., 1999), and hypertension (Smyth et al., 2008), indicating that LPs are also important signalling molecules under pathological condition (Gardell et al., 2006). Accordingly, the development of pharmacological tools that interfere with receptor-mediated LP signalling is nowadays thought to represent a promising approach for the development of novel therapeutic strategies against vascular diseases (Gardell et al., 2006).
One of the less examined LPs is lysophosphatidylinositol (LPI), a glycerol-based LP that has recently been reported to specifically interact with the G-protein-coupled receptor 55 (GPR55) (Oka et al., 2007; Waldeck-Weiermair et al., 2008) that mediates a transient Ca2+ increase upon stimulation of endothelial cells with LPI (Henstridge et al., 2008; Lauckner et al., 2008; Waldeck-Weiermair et al., 2008). This observation indicates that vascular endothelial cells are targets of LPI and might further point to LPI as an important modulator of endothelial functions under physiological and pathological conditions. Notably, GPR55 activation resulted in the suppression of a reconstituted neuronal M-type K+ current in GPR55 transfected HEK 293 cells (Lauckner et al., 2008), suggesting that GPR55 stimulation might have multiple consequences for the regulation of the membrane potential. It has been reported that several LPs such as lysophosphatidylcholin (LPC) exert significant effects on plasma membrane currents of endothelial as well as smooth muscle cells (Terasawa et al., 2002; Kuhlmannm et al., 2004) suggesting that LPs are capable of modulating vascular reactivity, predominantly by affecting ionic homeostasis in vascular cells. In vascular endothelial cells, the membrane potential is crucially important for various cell functions, including Ca2+ entry (Nilius and Droogmans, 2001), nitric oxide formation and the generation of endothelium-derived hyperpolarizing factor (EDHF; Chen and Suzuki, 1990; Graier et al., 1996), and free-radical production (McCarty, 1999). Moreover, changes in endothelial membrane potential in situ via myo-endothelial gap junctions influence the membrane potential of underlying smooth muscle cells (Beny and Pacicca, 1994) and, hence, have profound influence on vascular tone.
Because little is known about the effects of LPI as a possible vascular signalling mediator on endothelial membrane potential, this study was designed to investigate the effects of LPI on intracellular Ca2+ concentration, membrane potential, and to explore the underlying ion conductance in endothelial cells.
The human umbilical vein derived endothelial cell line, EA.hy926 (Edgell et al., 1983) at passage >45 was grown in DMEM containing 10% FCS and 1% HAT (5 mM hypoxanthine, 20 µM aminopterin, 0.8 mM thymidine) and cells were maintained in an incubator at 37°C in 5% CO2 atmosphere. Cells were plated on 30 mm glass cover slips at least 2 days before use in experiments (Paltauf-Doburzynska et al., 2000).
Adherent cells of approximately 70–80% confluence were transiently transfected with 2–3 µg cDNA of the respective vectors using TransFast™ Transfection Reagent (Promega, Mannheim, Germany) according to the instruction manual and as previously described (Trenker et al., 2007). Cells were used approximately 36 h after transfection. For the knock-down of GPR55 small interference RNA (siRNA) (Sense sequence of GPR55 siRNA: 5′-GGUUCUUGGCCAUCCGUUAtt-3′, Qiagen, Cambridge, MA) was used. Cells were transfected with either the siRNAs or the AllStars® negative control siRNA (Cat. No. 1027281, Qiagen, Cambridge, MA) as described previously (Trenker et al., 2007; Waldeck-Weiermair et al., 2008). Efficiency of the siRNA used was positively tested in our previous work (Waldeck-Weiermair et al., 2008). For GPR55, overexpression cells were transiently transfected with a vector encoding GPR55 as previously described (Waldeck-Weiermair et al., 2008).
Cytosolic free-Ca2+ was measured using Fura-2/AM as previously described (Graier et al., 1996; Paltauf-Doburzynska et al., 2000). Briefly, cells were loaded with 2 µM fura-2/AM for 45 min at room temperature. Prior to experiments, cells were washed and equilibrated for a further 20 min. Subsequently, cells were illuminated on an inverted microscope (Eclipse 300 TE, Nikon, Vienna) alternatively at 340 and 380 nm (filters: 340HTI15 and 380HTI15; Omega Optical Brattleboro, VT, USA) and emitted light was collected at 510 nm (510WB40 emission; Omega Optical) using a cooled charge-coupled device camera (−30°C; Quantix KAF 1400G2, Roper Scientific, Acton, MA, USA). All Ca2+ measurements were performed with a 40× 1.3 N.A. oil-immersion objective (Plan Fluor, Nikon, Vienna) (Paltauf-Doburzynska et al., 1998; Frieden and Graier, 2000).
Whole cell membrane currents and membrane potential were recorded using the perforated patch-clamp technique (Frieden and Graier, 2000; Bondarenko, 2004; Bondarenko and Sagach, 2006). The conventional whole cell technique was used only in those experiments in which cells were dialysed with Cs+ -based solution. For membrane perforation, amphothericin B (300 µM) was included into the pipette solution. Single channel activity of large conductance, Ca2+-activated, K+ channels (BKCa channels; nomenclature follows Alexander et al., 2009) was recorded in the cell-attached mode (Malli et al., 2003). Membrane currents and potential were recorded using a List EPC7 amplifier (List, Germany). Borosilicate glass pipettes were pulled with a Narishige puller (Narishige Co. Ltd, Tokyo, Japan), fire-polished and had a resistance of 4–6 MΩ. The signals obtained were low pass filtered at 1 kHz, and digitized with a sample rate of 10 kHz using a Digidata 1200A A/D converter (Axon Instruments, Foster City, CA, USA). Data collection and analysis were performed using Clampex and Clampfit software of pClamp (version 8.2, Axon Instruments). The holding potential in whole cell experiments was −40 mV. Voltage ramps of 1 s duration from −80 mV to +80 mV were delivered every 5 s.
The standard external solution contained (in mM): 145 NaCl, 5 KCl, 1.2 MgCl2, 10 HEPES, 10 glucose, 2.4 CaCl2 (extracellular buffer #1, EB#1). In Ca2+-free solutions, MgCl2 was increased to 2.2 mM and 1 mM EGTA was added. Patch pipettes were filled with a solution containing (in mM): 100 potassium aspartate (KAsp), 40 KCl, 10 HEPES, 2 MgCl2, 0,2 EGTA (pipette solution #1, PS#1). For studying the Na+-dependency of LPI-induced responses, bath solution contained: 150 NaCl, 1,2 MgCl2, 10 glucose, 10 HEPES, 1 EGTA (extracellular buffer #2, EB#2) and after current development in response to LPI, Na+ was omitted by substituting with choline. In these experiments a Cs+-based pipette solution contained the following components (in mM): 100 CsMeSO3, 40 CsCl, 10 HEPES, 2 MgCl2, 10 EGTA (pipette solution #2, PS#2). For single channel recordings of BKCa channels in the cell-attached mode, the bath solution contained (in mM): 150 KCl, 10 HEPES, 10 glucose, 5 EGTA and 4931 mM CaCl2 to set free Ca2+ concentration at 10 µM (calculated by the programme CaBuf, G. Droogmans, Leuven, Belgium; ftp://ftp.cc.kuleuven.ac.be/pub/droogmans/cabuf.zip) (extracellular buffer #3, EB#3) Pipette solution contained (mM): 100 KAsp, 40 KCl, 10 HEPES, 2 MgCl2, 5 EGTA and 1.924 CaCl2 to set free Ca2+ concentration to 100 nM (pipette solution #3, PS#3). Pipette solution in inside-out experiments for the activation of non-selective cation channels contained (in mM): 140 NaCl, 5 KCl, 2,4 CaCl2, 1,2 MgCl2, 10 HEPES and 10 d-glucose (pipette solution #4, PS#4). The respective bath solution contained (in mM): 140 CsCl, 1 MgCl2, 10 HEPES, 5 EGTA and free Ca2+ concentration was set to 100 nM by adding 1.924 CaCl2 (extracellular buffer #4, EB#4). The pH of the pipette solutions was adjusted to 7.25 with either KOH or CsOH. The pH of the external solutions was adjusted to 7.40 with either NaOH or CsOH (for Na+-free solution).
Cytosolic free Na+ was measured using CoroNa™ Green (Meier et al., 2006) following the standard procedure (Ohta et al., 2008). Briefly, cells were loaded with 5 µM CoroNa™ Green/AM for 30 min at room temperature followed by an equilibration period of 20 min. After washing the cells three times, cells were illuminated on an inverted microscope at 433 nm (433DF15, Omega Optical) and emission was collected at 535 nm (535AF26, Omega Optical, Brattleboro, USA). Changes in cytosolic Na+ were expressed as ΔF430/F0.
Analysis of variance was performed and statistical significance was verified using Scheffe's post hoc F-test. Level of significance was defined as P < 0.05.
Fura-2/AM and CoroNa™ Green/AM, gramicidin and cell culture chemicals were obtained from Invitrogen (Vienna, Austria). Fetal bovine serum was from PPA Laboratories (Linz, Austria). LPI, Dulbecco's modified Eagle's medium (DMEM) and all other chemicals were purchased from Sigma (Vienna, Austria).
In the presence of extracellular Ca2+, cell stimulation with 5 µM LPI induced a transient rise in cytosolic free [Ca2+], which returned to the basal level within 2–4 min even in the presence of 2 mM extracellular Ca2+ (Figure 1A). The comparison of LPI-induced Ca2+ signalling in the presence of extracellular Ca2+ with its effect in nominal Ca2+-free solution (Figure 1B) indicated that LPI mainly mobilized Ca2+ from internal Ca2+ stores, whereas Ca2+ entry contributed only marginally to the cytsolic Ca2+ elevation in this early phase while the sustained Ca2+ rise reflected Ca2+ entry. The concentration-response analysis in respect of cytosolic Ca2+ elevation in response to LPI revealed the initial intracellular Ca2+ mobilization to be more sensitive than the sustained Ca2+ entry (Figure 1C).
The initial cytosolic Ca2+ elevation upon LPI in the presence of extracellular Ca2+ was accompanied by a transient hyperpolarization that reached maximal amplitude of 11.4 ± 1.7 mV (n = 9) within 100 s. Following the initial hyperpolarization, a slowly developing sustained depolarization of 20.1 ± 2.5 mV (n = 9) above the resting membrane potential occurred within 250–300 s (Figure 1D). The concentration-response analyses revealed similar sensitivities of the initial hyperpolarization and subsequent depolarization (Figure 1E) compared with the respective Ca2+ signals (Figure 1C).
Upon repetitive applications, the LPI-induced initial hyperpolarization was markedly reduced or absent while the sustained depolarization remained unchanged (Figure 1F). In agreement with these findings, LPI failed to initiate repetitively the respective outward current that accompanied membrane hyperpolarization upon the first stimulation while a sustained inward current always occurred upon any LPI stimulation (Figure 1G).
Because in the cell model used, GPR55 was found to be constitutively expressed and to account for a fast, transient Ca2+ elevation upon stimulation with LPI (Waldeck-Weiermair et al., 2008), the contribution of this receptor to LPI-induced Ca2+ signalling and electrical responses in endothelial cells was further explored. In agreement with these findings, the GPR55 blocker rimonabant (1 µM) (Lauckner et al., 2008; Waldeck-Weiermair et al., 2008) diminished the LPI-initiated immediate Ca2+ transient (Figure 2A). Moreover, rimonabant (1 µM) prevented the transient membrane hyperpolarization (Figure 2B,C) and outward current in response to 5 µM LPI (Figure 2D). In contrast, the sustained membrane depolarization (Figure 2B,C) and inward current (Figure 2B) in response to LPI were not affected by the GPR55 blocker.
However, there is some controversy over to the pharmacological potential of putative inhibitors/activators of GPR55 (see Ryberg et al., 2007; Henstridge et al., 2008; Waldeck-Weiermair et al., 2008; Kapur et al., 2009; Ross, 2009; Yin et al., 2009). So, the involvement of GPR55 was further investigated by altering expression of GPR55. In line with these findings and our previous work using higher LPI concentrations (Waldeck-Weiermair et al., 2008), intracellular Ca2+ release triggered by LPI (5 µM) was strongly enhanced in cells that over-expressed GPR55 (Figure 3A) while this response was strongly attenuated in cells treated with siRNA against GPR55 (Figure 3B). Moreover, the initial membrane hyperpolarization in response to LPI was reduced in siRNA treated cells and augmented when GPR55 was over-expressed (Figure 3C). In contrast, the subsequent membrane depolarization remained unaffected [Control: 10.50 ± 2.10 mV (n = 4); GPR55 over-expressing cells: 12.16 ± 2.20 mv (n = 6)].
To further assess the mechanistic differences between LPI-triggered initial hyperpolarization and the long-lasting membrane depolarization, the effect of the phospholipase C inhibitor U73122 (Figure 4A) and of the G protein inhibitor GDPβS (Figure 4B) was tested in conventional patch clamp experiments. U73122 as well as GDPβS prevented LPI- (3 µM) induced membrane hyperpolarization while the depolarizing effect of LPI remained unaffected in the presence of GDPβS. Moreover, in inside-out experiments LPI in the bath activated a non-selective ion current (Figure 4C), thus, indicating that LPI, besides its stimulatory activity at GPR55, might act independently from a receptor directly on plasma membrane ion channels.
These data indicate that the initial and transient ionic responses to LPI essentially depend on GPR55, while the subsequent, slowly developing, sustained electrical responses appear to be independent of this orphan receptor and mediated via direct effects on non-selective cation channels in endothelial cells.
To describe the transmembrane currents that underlie the LPI-induced changes in membrane potential, experiments in the voltage clamp mode were performed. When membrane voltage was kept at −40 mV, which is close to the resting membrane potential (−40.9 ± 2.2 mV; n = 23) of this cell type, 5 µM LPI induced a biphasic response that showed a transient outward current (Hyperpolarization phase) followed by a sustained inward current (Depolarization phase) (EB#1, PS#1) (Figures 1G and and5A5A).
Voltage ramps between −80 to +80 mV during the hyperpolarization (outward current) phase (Figure 5A) revealed a current potential curve with the common characteristics of an agonist-activated current, mainly achieved by BKCa channels in whole cell recordings of these endothelial cells (Figure 5B) (Frieden and Graier, 2000). In agreement with this assumption, LPI induced rapid activation of large conductance channels in the cell-attached configuration (Figure 5C). In the whole-cell configuration, the outward current did not appear when pipettes were filled with Ca2+-free Cs+-containing solution (PS#2) in order to avoid KCa currents (data not shown). Moreover, membrane hyperpolarization to LPI was sensitive to charybdotoxin (Figure 5D). Charybdotoxin (100 nM) produced a gradual endothelial depolarization (Figure 5D) that points to the contribution of charybdotoxin-sensitive KCa channels to the resting membrane potential (Ledoux et al., 2008). Altogether, these results indicate that the LPI-induced GPR55-dependent hyperpolarization is due to an activation of large conductance, Ca2+-dependent, charybdotoxin-sensitive, K+ channels.
Next, the sustained inward current that accounts for the slowly developing membrane depolarization in endothelial cells in response to LPI was investigated. Using voltage ramps between −80 to +80 mV in the inward current phase (Figure 6A) revealed a linear current-potential curve (Figure 6B). We intended to characterize the nature of the current(s) that was/were responsible for the LPI-induced membrane depolarization in endothelial cells. First, the involvement of Cl- channels in LPI-induced depolarization was tested using the Cl- channel blocker 4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (DIDS). DIDS (100 µM and 1 mM) failed to prevent endothelial depolarization (Figure 6C) evoked by 5 µM LPI, thus, suggesting that Cl- channels are unlikely to be involved in the LPI-induced electrical responses in endothelial cells.
To evaluate the role of Ca2+ entry in the LPI-induced electrical responses, LPI was added either in nominal Ca2+-free or Ni2+-containing solution. Ca2+ withdrawal evoked a gradual depolarization with the amplitude of 11.4 ± 2.1 mV (n = 6). In the absence of extracellular Ca2+, LPI produced a more pronounced initial hyperpolarization with the mean amplitude of −15.4 ± 2.7 mV (n = 5) (Figure 7A). However, the subsequent depolarization in Ca2+-free solution was reduced by about 42 ± 4% (n = 5) compared with the depolarizing effect of LPI in Ca2+ containing buffer (Figure 7A,F). Ca2+ re-addition in the continued presence of LPI produced a hyperpolarization followed by a depolarization above the level attained in the presence of LPI in Ca2+-free solution (Figure 7B,F). Similar results were obtained if Ca2+ entry was prevented by 2 mM Ni2+ (Figure 7C,F). Addition of Ni2+ caused a gradual depolarization pointing to a basal Ca2+ entry as a determinant of the resting membrane potential in this cell type. Subsequent addition of 5 µM LPI evoked a slightly increased amplitude of the transient hyperpolarization (18.7 ± 2.4 mV, n = 7) followed by a sustained depolarization that was significantly reduced by 51 ± 4% compared with the corresponding experiments that were performed in the absence of Ni2+ (Figure 7C,F). In line with these findings, the LPI-induced inward current (repetitive stimulation) was reduced upon addition of 2 mM Ni2+ into the bath by 50 ± 6% (Figure 7D). Next, we examined the sensitivity of the depolarization phase to La3+, a known blocker of store-operated and non-selective cation channels (Nilius and Droogmans, 2001). In these experiments, an addition of 100 µM La3+ during the sustained depolarization upon a repetitive LPI-administration (5 µM) reduced membrane depolarization by 53 ± 7% (Figure 7E,F).
These results indicate that the lack of Ca2+ entry slightly facilitates LPI-induced hyperpolarization and reduces the subsequent sustained depolarization phase, indicating that Ca2+ entry partially (approximately 50%) accounts for the depolarizing effect of LPI. Together with the current-potential relationship of the current that is associated with the sustained depolarization phase upon LPI stimulation (Figure 6B), these results point to the involvement of Ca2+ permeable, non-selective channels, activated by LPI action and contributing to LPI-induced endothelial cell depolarization.
In order to study whether or not Na+ entry via such non-selective cation channels also contributes to the LPI-induced depolarization, experiments in the absence of extracellular Na+ (Na+ substituted for choline) and nominal Ca2+-free buffer were performed. Under such conditions, LPI-induced sustained depolarization was reduced by 94 ± 3% compared with the respective control (Figure 8A). In voltage-clamp experiments (PS#2, holding potential −40 mV, voltage ramps from −80 to +80 mV), LPI-induced inward current was abolished upon substitution of extracellular Na+ with equimolar choline (Figure 8B). These data suggest that Ca2+- and Na+-permeable non-selective cation channels account for the slowly developing but sustained depolarization upon stimulation of LPI.
In line with the findings above, LPI resulted in a slow accumulation of intracellular Na+ (Figure 8C), This effect was independent from GPR55 as indicated by the lack of an effect of GPR55 knock-down on LPI-triggered cytosolic Na+ elevation (Figure 8C). The concentration dependency that indicated a great difference in the effects initiated by 5 versus 10 µM LPI (Figure 8D, left) was similar to that obtained for Ca2+ entry/plateau phase (Figure 1C) and membrane depolarization (Figure 1E). In the absence of extracellular Na+, LPI failed to increase cytosolic Na+ levels (Figure 8D, right).
Because in our experiments above, LPI-induced sustained membrane depolarization dependent on Ca2+/Na+ inward currents (Figures 6–8), further experiments were performed in order to assess the possible involvement of the Na/K-ATPase in LPI-induced depolarization. First, 5 µM LPI prevented, in a reversible manner, membrane hyperpolarization upon re-adddition of extracellular K+ (Figure 9A), a standard protocol that reveals Na/K-ATPase-dependent repolarization upon the re-addition of K+ (Bondarenko and Sagach, 2006). Furthermore, preincubation of endothelial cells with 250 µM ouabain, which yielded strong membrane depolarization, reduced LPI (5 µM)-induced sustained depolarization from 20.1 ± 2.5–8.5 ± 1.5 mV (n = 9) (Figure 9B,C). The subsequent inhibition of the non-selective cation channels by 100 µM La3+ in the presence of ouabain completely prevented LPI-induced, sustained membrane depolarization (Figure 9B,C).
The present study demonstrated that the endogenous agonist of GPR55, LPI induced a variety of electrical responses in endothelial cells. The immediate and rather transient Ca2+ elevation, and its associated K+ current and membrane hyperpolarization to LPI were mediated via GPR55. However, the slowly developing and sustained inward current that resulted in intracellular Na+ loading and membrane depolarization was independent of GPR55 and was due to inhibition of Na/K-ATPase and direct activation of non-selective cation channels. Thus, LPI is a potent signalling molecule affecting endothelial cells by modulating multiple electrical responses. Because, these effects are only partially linked to GPR55, the physiological meaning of LPI as an intercellular messenger within the vascular wall exceeds that of just an activator of GPR55. Accordingly, in the ongoing evaluation of the physiological role of GPR55 and its endogenous agonists, these biphasic responses need to be considered.
LPI recently received much attention following several concordant reports that identified LPI but not other lysophospholipids, such like lysophosphatidylcholine, lysophosphatidylethanolamine, lysophosphatidylserine or lysophosphatidic acid, as endogenous activators of GPR55 (Oka et al., 2007; 2008; Henstridge et al., 2008; Waldeck-Weiermair et al., 2008). Because the physiological concentrations of LPI were found to be 1–15 µM (Xiao et al., 2001), the LPI concentrations we used here are close to those found in vivo. There is an ongoing discussion whether or not GPR55 serves as receptor for endocannabinoids, such as anandamide and 2-arachidonyl glycerol, though it obviously shares some similar pharmacology with classical cannabinoid receptors (Ross, 2009), The two putative endogenous GPR55 agonists, LPI and anandamide, share certain biological activities such as stimulation of intracellular Ca2+ signalling (Howlett and Mukhopadhyay, 2000; Oka et al., 2007; Henstridge et al., 2008; Lauckner et al., 2008; Waldeck-Weiermair et al., 2008) and the initiation of mitogenic signals (Falasca and Corda, 1994; Corda et al., 2002; Waldeck-Weiermair et al., 2008). Regardless of their proliferative potential, anandamide and LPI predominantly exhibit inhibition of endothelial cell migration, tumour growth and angiogenesis (Murugesan and Fox, 1996; Flygare and Sander, 2008).
Endothelial cells express functional GPR55 (Waldeck-Weiermair et al., 2008) and produce its endogenous agonist, LPI (Hong et al., 1985; Kaya et al., 1989), making it more likely that LPI could be an intra-vascular messenger. To understand the physiological role of LPI, an in-depth understanding of the signalling cascades initiated by LPI in endothelial cells is essential. In this study, LPI was found to exhibit several autonomous effects in endothelial cells of which the immediate, rather transient signalling depended exclusively on GPR55. This conclusion is based on our findings that the antagonist of GPR55, rimonabant (Lauckner et al., 2008; Waldeck-Weiermair et al., 2008) and siRNA-mediated GPR55 knock-down prevented the rapid intracellular Ca2+ release in response to LPI, confirming our earlier results that LPI triggered intracellular Ca2+ mobilization via GPR55 (Waldeck-Weiermair et al., 2008). Moreover, our findings that the inhibition/knock-down of GPR55 prevented LPI-induced transient membrane hyperpolarization and outward current while GPR55 over-expression strongly enhanced LPI-induced hyperpolarization, further supports our assumption that the initial GPR55-mediated Ca2+ elevation/release is accompanied by activation of plasma membrane ion channels, that were characterized as intermediate and large conductance Ca2+-activated K+ channels, the main carrier for Ca2+-induced membrane hyperpolarization in this cell type (Frieden and Graier, 2000; Malli et al., 2003). In contrast to other G protein-coupled receptor agonists in this cell type (Malli et al., 2007), LPI initiated a transient Ca2+ elevation even in the presence of extracellular Ca2+, which points to a limited activity of store-operated Ca2+ entry (SOCE). Nevertheless, LPI also induced inward currents that were not linked to increased cytosolic Ca2+. These findings were in apparent contradiction to reports describing LPI as direct activator of SOCE in smooth muscle cells (Smani et al., 2003; 2007; Singaravelu et al., 2006) and the hypothesis that LPI represents a Ca2+-independent, phospholipase A2-derived, mediator for activation of SOCE (Smani et al., 2004). These differences might be due to the expression of a wide variety of Ca2+-permeable ion channels in different cell types. Moreover, because LPI-triggered sustained membrane depolarization and inward current were not affected by inhibition of GPR55 by rimonabant or the molecular manipulation of GPR55 expression, LPI obviously exhibited its sustained effects independently of this orphan receptor. This assumption was further supported by our findings that LPI-induced slow cellular Na+ loading was independent of the expression level of GPR55.
Accordingly, the LPI-induced biphasic responses of the membrane potential differs from the reported monophasic hyperpolarization elicited by LPC in endothelial cells (Erdogan et al., 2007) and coronary artery smooth muscle cells (Terasawa et al., 2002), which respond by a slowly developing depolarization due to activation of a non-selective cation current. However, the biphasic effect of LPI on electrical responses reported here resembles the effect of LPC on renal arterial smooth muscle cells (Jabr et al., 2000).
To investigate the nature of the GPR55-independent inward current, blockers of several endothelial transporters and channels were tested. As LPI-evoked depolarization was not affected by the Cl- channel blocker DIDS, the involvement of Cl- channels (Nilius and Droogmans, 2001) in the membrane depolarization to LPI is unlikely.
In contrast, depletion of Ca2+ in the bath solution, which produced gradual depolarization possibly due to inhibition of basal Ca2+ entry and a subsequent reduction of BKCa channel activity, strongly reduced LPI-induced depolarization. Moreover, the Ca2+ entry blocker Ni2+ produced a gradual depolarization and diminished LPI-induced depolarization, thus, indicating that LPI-induced sustained depolarization requires Ca2+ entry. We also used La3+, which selectively inhibits SOCE and Ca2+ permeable non-selective cationic channels (Nilius and Droogmans, 2001; Terasawa et al., 2002) at low concentrations. Our findings that La3+ attenuated the LPI-induced depolarization further support our assumption of the importance of Ca2+ entry for LPI-induced depolarization and point to a GPR55-independent activation of Ca2+ permeable non-selective cationic channel(s) by LPI.
Our experiments further showed that LPI increased also the Na+ permeability of endothelial cells. After cell dialysis with Cs+- containing solution, LPI evoked a linear current that reversed at 0 mV, a property of non-selective cationic current (Jabr et al., 2000; Terasawa et al., 2002). Na+ substitution with choline inhibited the current, indicating a significant Na+ permeability stimulated by LPI, which contributed to membrane depolarization. Furthermore, a direct slow and GPR55-independent Na+ loading was observed in response to LPI, indicating that LPI, in a GPR55-independent manner, induces a Ca2+/Na+-permeable non-selective cation channel in endothelial cells that favours Na+ to enter the cell and, thus, to initiate depolarization in the absence of an increase in cytosolic free Ca2+ concentration.
Removal of extracellular Ca2+ or the application of inhibitors of (non-selective) Ca2+-permeable cation channels (Ni2+, La3+) only prevented about 50% of the depolarization to LPI. These data point to an additional GPR55-independent effect of LPI besides the activation of a non-selective cation channel. Because Ni2+ is also an inhibitor of the plasma membrane Na+/Ca2+ exchanger (Iwamoto et al., 1999), this carrier can be excluded. Notably, LPI failed to trigger membrane depolarization in the absence of Na+, indicating that the remaining membrane depolarization upon LPI relies on Na+. Because other LPs including LPC and sphingosine exhibit inhibitory properties on the Na/K-ATPase (Oishi et al., 1990), we also evaluated the Na/K-ATPase as a putative target of LPI. Implementation of K+ re-addition protocols (Bondarenko and Sagach, 2006) demonstrated that, in the presence of LPI, a K+ -induced hyperpolarization was strongly reduced, indicating that LPI inhibited the Na/K-ATPase. In line with these findings, LPI-induced Ni2+/La3+-insensitive depolarization was prevented by ouabain, an inhibitor of the Na/K-ATPase. Moreover, only the combination of Ni2+/La3+ and ouabain completely prevented the sustained deplolarization induced by LPI.
Although the function of GPR55 as a common binding site for LPI and anandamide was only demonstrated in regard to inositol 1,4,5-trisphosphate (IP3)- dependent intracellular Ca2+ signalling (Oka et al., 2007; Lauckner et al., 2008; Waldeck-Weiermair et al., 2008), it is tempting to speculate that several other physiological effects of both compounds are due to their common action on GPR55. However, these compounds considerably differ in terms of their GPR55-independent potential to affect distinct ion channels and carriers and their role in the control of blood vessel tone. In particular, LPI was described as a phospholipase A2-derived messenger for the activation of SOCE (Smani et al., 2003; 2007; Singaravelu et al., 2006), and to activate non-selective cation channels and to prevent Na/K-ATPase activity (this study). Anandamide was described to activate arachidonic acid-regulated Ca2+-selective channels (Shuttleworth et al., 2004) and the transient receptor potential cation channel, TRPV1 (Zygmunt et al., 1999), and to trigger formation of 5,6-epoxyeicosatrienoic acid as a messenger for Ca2+-influx (Graier et al., 1995; Hoebel et al., 1997). Thus, the differences in the physiological effects of these two putative endogenous GPR55 agonists are likely to be due to their different repertoire of GPR55-independent effects. Our present data on LPI-induced depolarization might explain the discrepancy between anandamide and LPI in their potential to control endothelium-dependent blood vessel tone: anandamide may act as an EDHF per se or be involved in its formation/release (see Randall and Kendall, 1998), while LPI inhibits acetylcholine-induced, endothelium-dependent, hyperpolarization of rat mesenteric arteries (Fukao et al., 1995).
In view of the reported more than 10-fold elevation of LPI in cancer-derived ascites fluid (Xiao et al., 2001), our data presented here deserve further attention. In particular, our findings that the GPR55-dependent effects of LPI tend to be more sensitive, compared with the GPR55-independent LPI effects, suggest a considerable shift in the biological activity of LPI, under certain conditions. However, this important aspect requires more extensive evaluation and exceeds the scope of the present work.
In conclusion, our results indicate that LPI exhibits GPR55-dependent and GPR55-independent effects in endothelial cells. Because endothelial functions critically depend on Ca2+ entry and membrane potential fluctuations, modulation of these parameters by LPI points to a unique messenger function of LPI in the vasculature.
The authors thank Mrs Anna Schreilechner, BS, for her excellent technical assistance, Dr C.J.S. Edgell (University of North Carolina, Chapel Hill, NC, UCA) for the EA.hy926 cells and Dr M. Frieden (University of Geneva, Geneva, Switzerland) for critically reviewing this manuscript. This work was supported by the Austrian Funds (F3010-B05, P20181-B05 and P21857-B18).
Shamim Naghdi is funded by the FWF P20181-B05 within the PhD programme Molecular Medicine of the Medical University of Graz.
None to declare.
Supporting Information: Teaching Materials; Figs 1–9 as PowerPoint slide.