Block of TRPC5 channels by 2-aminoethoxydiphenyl borate: a differential, extracellular and voltage-dependent effect

Introduction

The transient receptor potential (TRP) superfamily comprises mammalian genes encoding calcium- and sodium-permeable channels that are divided into six subfamilies: TRPC (canonical), TRPM (melastatin), TRPV (vanilloid), TRPP (polycystin), TRPML (mucolipin) and TRPA (ANKTM1) (Clapham, 2003). The channels have diverse functions including involvement in Ca²⁺ store-operated and G-protein coupled receptor pathways, cell volume regulation, and pheromone, redox, taste and temperature sensation. The TRPC family has seven members, all of which have been associated with Ca²⁺-permeable ion channel elements of store- and receptor-operated plasma membrane signalling. Based on amino-acid sequence they are subdivided into four groups (TRPC1, TRPC2, TRPC4/5 and TRPC3/6/7). Apart from TRPC2, which is not thought to be expressed in human, TRPC expression may be relevant to a range of human cells in the nervous system, blood vessels and blood cells, gastrointestinal tract and kidney (Sossey-Alaoui et al., 1999; Vannier et al., 1999; Xu & Beech, 2001; Riccio et al., 2002; Zitt et al., 2002; Clapham 2003; Beech et al., 2004). However, the physiological function and pharmacological regulation of TRPC channels are still unclear.

2-Aminoethoxydiphenyl borate (2-APB) was originally introduced as a membrane-permeable inhibitor of IP₃ receptors (IP₃Rs) and has since been widely used to examine the functions of these receptors and the functions of other Ca²⁺ signalling mechanisms such as store-operated Ca²⁺-entry (Bootman et al., 2002). Commonly, 75 μM 2-APB is sufficient to strongly inhibit Ca²⁺ influx via store-operated channels (SOCs) in native cell types (Bootman et al., 2002; Ma et al., 2002; Flemming et al., 2003). Furthermore, micromolar 2-APB has effects on heterologously expressed TRP channels, such as TRPC1 (Delmas et al., 2002), TRPC3 (Ma et al., 2000; Trebak et al., 2002), TRPC5 (Lee et al., 2003b), TRPV6 (Voets et al., 2001; Schindl et al., 2002), TRPM8 (Hu et al., 2004) and TRPV1-3 (Chung et al., 2004; Hu et al., 2004).

In this study, we explore the mechanism of the effect of 2-APB on human TRPC5 and make comparisons with TRPC6, TRPM2 and TRPM3.

Methods

For studies of human TRPC5 (Accession Number AF054568) and TRPM2 (NM_003307) HEK-293 cells had stable tetracycline (Tet)-regulated expression of cDNA encoding TRPC5 (Zeng et al., 2004) or TRPM2 (McHugh et al., 2003). Cells were grown in DMEM-F12 (Gibco, U.K.) medium containing 10% fetal calf serum, 100 U ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin. Cells were maintained at 37°C under 95% air and 5% CO₂ and replated on coverslips prior to experiments. The expression of TRPC5 and TRPM2 was induced by 1 μg ml⁻¹ tetracycline (Tet+) for 24–72 h before recording. Noninduced cells without addition of tetracycline (Tet−) were used as control. HEK-293 cells stably expressing mouse TRPC6 and transiently expressing human TRPM3 have been described (Boulay, 2002; Grimm et al., 2003).

Voltage clamp was performed at room temperature (23–26°C) with the whole-cell, inside-out and outside-out patch configurations. Signals were amplified with an Axopatch 200A patch clamp amplifier and controlled with pClamp software 6.0 (Axon). A 1-s ramp voltage protocol from −100 to +100 mV was applied at a frequency of 0.1 Hz and from a holding potential of −60 mV. Signals were digitally sampled at 3 kHz after analogue filtering with a cutoff frequency of 1 kHz. For Figure 5c, data were smoothed by adjacent five-point averaging to reduce noise prior to calculation of the percentage block. For inside-out or outside-out patches and voltage dependence studies, step voltages from −100 to +90 mV were applied and signals sampled at 10 kHz. Patch pipettes were made from borosilicate glass capillary tubing with an outside diameter of 1 mm and inside diameter of 0.58 mm (Clark Electromedical Instruments, Reading, U.K.). After fire-polishing and filling with pipette solution, the resistance of pipettes was 3–5 MΩ. The standard patch pipette solution contained (mM): 135 CsCl, 1 EGTA, 2 MgCl₂, 10 HEPES, 5 sodium ATP, 0.1 sodium GTP, pH titrated to 7.2 with CsOH, and osmolarity adjusted to 290 mOsm with mannitol. For some recordings, as indicated, a pipette solution containing 200 nM free Ca²⁺ was used (mM): 115 CsCl, 10 EGTA, 2 MgCl₂, 10 HEPES, 5 sodium ATP, 5.7 CaCl₂, pH titrated to 7.2 with CsOH, and osmolarity adjusted to 290 mOsm with mannitol. To study the voltage dependence of block, the 200 nM Ca²⁺ pipette solution was used, but with low chloride (120 CsOH, 120 L-glutamic acid, 10 EGTA, 2 MgCl₂, 10 HEPES, 5.7 CaCl₂, pH 7.2, osmolarity 290 mOsm). For TRPM2 experiments, 0.5 mM ADP-ribose was included in the 200 nM Ca²⁺ pipette solution. The standard bath solution contained (mM): 130 NaCl, 5 KCl, 8 D-glucose, 10 HEPES, 1.2 MgCl₂ and 1.5 CaCl₂, pH titrated to 7.4 with NaOH. For inside-out patch experiments the bath solution contained (mM): 130 CsCl, 1 EGTA, 2 MgCl₂, 10 HEPES, 5 sodium ATP, titrated to pH 7.2 with CsOH; The standard bath solution was in the pipette. The recording chamber had a volume of 150 μl and was perfused at a rate of about 2 ml min⁻¹.

Ca²⁺ imaging was previously described (Xu & Beech, 2001). Cells were preincubated with 1 μM fura PE3-AM at 37°C for 1 h in standard bath solution, followed by a 15–30 min period in standard bath solution at room temperature. The ratio of fura-PE3 fluorescence excited alternately at 340 and 380 nm was monitored (Improvision, U.K.) and emission was collected via a 510-nm filter and sampled by a CCD camera (Orca ER; Hamamatsu, Japan). Images were sampled every 10 s in pairs for the two excitation wavelengths and analysed off-line using regions of interest to select individual cells. Imaging was controlled by Openlab 2.0 software (Image Processing & Vision Company Ltd, U.K.). In some cases, store-depletion was achieved by incubating cells for 30 min with 1 μM thapsigargin (TG). TRPM3 methods were as described (Grimm et al., 2003). In brief, fluorescence quenching by Mn²⁺ entry was studied using the fura-2 isosbestic excitation wavelength of 360 nm. Measurements were made 200 s after application of 1 mM Mn²⁺ in the presence or absence of 100 μM 2-APB. In all experiments, transfected cells were identified by their YFP-fluorescence at an excitation wavelength of 480 nm. All the experiments were at room temperature.

The 5-bromo-2′ deoxy-uridine (BRDU) labelling kit (Roche) was used for cell proliferation measurement. Tet-induced TRPC5 cells were split onto a 24-well plate and cultured with test substance for 36–72 h and then incubated with BRDU for 2 h. After removal of labelling medium, cells were washed, fixed and incubated with anti-BRDU antibody (1 : 1000) for 2 h at room temperature. Cells were washed three times and added substrate solution for colour development and absorbance measurement.

All salts and reagents were from Sigma (U.K.) or BDH (British Drug House, Poole, U.K.). 2-APB, TG, carbachol (CCH), methoxyverapamil (D600), nifedipine, N-propargylnitrendipine (MRS 1845), gadolinium chloride (Gd³⁺) and bovine serum albumin (BSA) were purchased from Sigma. Berberine was from Northeast Pharmaceuticals (China). Fura-PE3 AM was from Molecular Probes. TG (10 mM), fura-PE3 AM (1 mM) and 2-APB (75 mM) were made up as stock solutions (stock concentrations are in parentheses) in 100% dimethyl sulphoxide (DMSO). DMSO (0.1%) had no effect on TRPC5 ionic current (n=5; e.g. Figure 1c).

Data are expressed as mean±s.e.m., where n is the number of individual experiments for electrophysiological recordings and cell proliferation analysis, and the number of cells from at least three independent experiments (one per coverslip) for Ca²⁺ imaging. Data sets were compared using unpaired Student's t-test with significance indicated if P<0.05. The concentration of half-maximal inhibition (IC₅₀) of 2-APB was given by curve fitting with the Hill equation.

Results

The Tet-induced expression of human TRPC5 in HEK293 cells was confirmed by Western blotting (Zeng et al., 2004). In Ca²⁺-imaging experiments, CCH acting at endogenous muscarinic receptors evoked a Ca²⁺_i signal in store-depleted cells only if TRPC5 expression was induced (Tet+, n=60–62) (Figure 1a). The effect was enhanced by application of gadolinium (Gd³⁺, 10 μM) and again there was no effect without TRPC5 (Tet−) (Figure 1a). In whole-cell recordings, CCH activation of TRPC5 was evident as an outward current at +80 mV and inward current at −80 mV (n=4) (Figure 1b). The reversal potential was close to 0 mV. CCH had no effect in noninduced cells (n=4, data not shown). The current–voltage relationship (IV) was similar to that described for mouse TRPC5 (Jung et al., 2003), showing inward and outward (‘double') rectification, giving something that is roughly ‘N-shaped' (Figure 1b, inset). In patch-clamp recordings, a similar current was activated directly by Gd³⁺ (10 μM) in Tet+ cells, whereas there was a small inhibitory effect of Gd³⁺ in noninduced cells (data not shown).

Bath-applied 2-APB essentially abolished TRPC5 current activated by Gd³⁺ (Figure 1c, d), or Gd³⁺ plus CCH (data not shown). The effect occurred promptly and was fully reversible on washout. It was concentration-dependent (IC₅₀ 19 μM) and the Hill coefficient was near to 1.0 (Figure 1e). No stimulatory effects were observed (2.8–100 μM 2-APB, n=6). In noninduced cells, current was small and had a linear IV, unlike that of TRPC5. 2-APB also inhibited this endogenous current but the effect was much smaller than in Tet+ cells (Figure 1f).

TRPC5 can be activated by store-depletion and so may be part of endogenous SOCs, some of which are inhibited by conventional Ca²⁺ channel blockers (Beech et al., 2004; Zeng et al., 2004). Effects of Ca²⁺ channel blockers on TRPC5 are also possible because there is general structural resemblance of TRP channels to voltage-gated Ca²⁺ channels. Therefore, we considered that established Ca²⁺ channel blockers might affect TRPC5 and possibly mimic the action of 2-APB. Measured by Ca²⁺-imaging, 10 μM methoxyverapamil (Figure 2a) or 1 μM nifedipine (Figure 2b) had no effect on TRPC5. Berberine (25 μM), an L- and a T-type Ca²⁺ current blocker (Xu et al., 1997), was tested by whole-cell recording and also had no effect (n=10, P>0.05; Figure 2c). N-propargylnitrendipine (MRS 1845, 10 or 100 μM), a derivative of 1,4-dihydropyridines and blocker of store-operated Ca²⁺ entry in HL-60 cells (Harper et al., 2003), had no effect on TRPC5 current (n=5, P>0.05; Figure 2d). Therefore, there is no functional site for conventional Ca²⁺ channel blockers on TRPC5, and the effect of 2-APB is not related to that of Ca²⁺ channel blockers.

2-APB may suppress function of human TRPC5 either because it inhibits IP₃Rs or because of some other potentially direct effect on the channel. In an effort to explore these possibilities we tested the effect of 2-APB on macroscopic TRPC5-mediated currents in excised membrane patches. In outside-out patch recordings from Tet-induced cells we demonstrated activity of TRPC5 stimulated by Gd³⁺. This Gd³⁺-induced current was abolished by bath-applied 2-APB (Figure 3a, b, n=3, P<0.001). Figure 3c shows typical original current traces for a range of voltages both before (control) and after application of 75 μM 2-APB. The noisy, macroscopic, activity of TRPC5 in the patch was inhibited by 2-APB. A similar blocking effect was observed in outside-out patches if the channel activator was La³⁺ instead of Gd³⁺ (data not shown).

In inside-out patches we demonstrated we had the correct membrane orientation as well as TRPC5 expression by bath-applying the stable GTP analogue GTP-γ-S (20 μM), which induced current (Figure 4a), as in studies of mouse TRPC5 (Jung et al., 2003). Bath-application of 2-APB to the intracellular face of the membrane had no effect on TRPC5 (Figure 4c, d). Gd³⁺ also had no stimulating effect when applied intracellularly (data not shown). Concerned that the critical intracellular binding proteins for 2-APB may be lost in excised patches, we also recorded TRPC5 current in whole-cell configuration, applying 2-APB at a high concentration (150 μM) to the inside of the cell via the patch pipette. Strikingly, a large TRPC5-mediated current remained 30 min after starting whole-cell recording (Figure 4d, n=3). Bath application of 2-APB at half the concentration (75 μM) abolished the TRPC5-mediated current (Figure 4d). Therefore, TRPC5 is resistant to intracellular 2-APB, but blocked by extracellular 2-APB.

Blockade of channels is commonly voltage-dependent, especially when the agent has a charged moiety in its structure, as 2-APB does. To explore if block of TRPC5 by 2-APB has voltage-dependence we preactivated TRPC5 with a lanthanide and applied 2-APB at 20 μM (Figure 5). We also checked the suitability of using a ramp voltage protocol by making comparisons with IVs constructed using square steps (Figure 5a, b). The original current traces during square voltage steps show no major time-dependent activation or inactivation events, although a modest decay of current is apparent when the voltage is initially stepped to −80 mV (Figure 5a). Nevertheless, despite this event and the obvious voltage-dependence of TRPC5 (i.e. the IV rectifies, Figure 5b), the ramp data align closely with square-step data (Figure 5b), suggesting any voltage-dependent switching events are rapid. Low Cl⁻ in the pipette solution was used to minimise the risk of contamination from endogenous Cl⁻ current and neither this (Figure 5b) nor the additional removal of Cl⁻ from the bath solution by replacement with glutamate (data not shown) changed the ‘N' shape IV that is characteristic of TRPC5 (cf Figures 1 and 2). The blocking effect of 2-APB on endogenous current in control cells shows little voltage-dependence, contrasting with inhibition of TRPC5 channels, which was weaker at positive membrane potentials (Figure 5c). Similar data were produced using a square-step voltage paradigm (75.8±1.7 and 53.8±7.1% block at −80 and +80 mV respectively, n=5) or with the ramp protocol and reduced extra- and intracellular Cl⁻ concentrations (74.2±0.7 and 59.8±3.0% block at −80 and +80 mV respectively, n=3). Therefore, blockade of TRPC5 by 2-APB has intrinsic voltage-dependence.

Differential effects of inhibitors on related ion channels can help indicate the physical location of the blocking site. Like the action on TRPC5, 2-APB at 75 μM abolished TRPC6 current. The onset of the effect was prompt and the block was readily removed on wash-out (Figure 6a). The reversal potential of TRPC6 current was near 0 mV and the IV showed outward rectification (Figure 6b) characteristic of TRPC6 (see review of TRPC6 data in Beech et al., 2004). A small effect of 2-APB was observed in control HEK-293 cells (Figure 6c, d), as in Tet− cells (see Figure 1f). The inhibitory effect of 2-APB also occurred in outside-out patch recordings from TRPC6 cells (n=3; e.g. Figure 6e, f). TRPM3 is a newly discovered member of the TRP family (Grimm et al., 2003; Lee et al., 2003a), and was also inhibited by 2-APB: Using an Mn²⁺ quenching protocol (Grimm et al., 2003), the tonic Mn²⁺ response in TRPM3-transfected cells was a 39.45±9.02% (three independent experiments, n>20 cells each) reduction in the F360 signal 200s after applying Mn²⁺. In the presence of 2-APB (100 μM) this was significantly reduced to 12.45±5.53% (P<0.05, three independent experiments), and thus to near the amplitude observed in nontransfected cells (10.59±2.1%, four independent experiments). TRPM2 is an important ADP-ribose and Ca²⁺-activated cationic channel (Perraud et al., 2001; McHugh et al., 2003). It is characterised by a linear IV, a reversal potential close to 0 mV, loss of inward current when Na⁺ is substituted by NMDG⁺, and resistance to Gd³⁺ (Figure 7a, c, d). 2-APB at 75 and 150 μM had no effect on TRPM2-mediated current but the effect on the small background current was evident in Tet− cells (Figure 7b–d).

To address the functional relevance of TRP channel block by 2-APB, we studied cell proliferation. A 36–72 h incubation with 75 μM 2-APB led to strikingly fewer HEK-293 cells compared with control (Figure 8a). This effect might simply be related to inhibition of IP₃R function. However, proliferation of DT40 cells that lack IP₃Rs was also suppressed by 2-APB (Figure 8a), suggesting other mechanisms are involved. To explore this further we determined the effect of 2-APB on BRDU incorporation as a measure of DNA synthesis (and thus cell proliferation) in cells with and without induction of TRPC5 expression. There was significantly greater effect when TRPC5 was expressed (Figure 8b), showing inhibition of cell proliferation by 2-APB is partly related to block of TRP channels.

Discussion

In this study we explored the blockade of TRP channels by 2-APB, focusing particularly on human TRPC5. 2-APB blocks the channel in a concentration-dependent manner and acts exclusively via a site that is accessible from the extracellular face of the membrane. This site has specificity for 2-APB because other Ca²⁺ channel blockers are ineffective. The lack of effect of intracellular 2-APB suggests functional IP₃Rs are not required for TRPC5 activation, or blockade by 2-APB. The blockade has an intriguing voltage-dependence, as if 2-APB binds to a membrane-spanning region of the channel within the electric field or modulates TRPC5-gating. Comparison with other TRP channels reveals similar sensitivity of TRPC6 and TRPM3, but resistance of TRPM2. Results from cell proliferation assays indicate 2-APB blockade of TRP channels may be functionally relevant.

Our data are consistent with lanthanides and 2-APB exerting their effects on TRPC5 via extracellular sites. The lanthanide effect is comparable to that described for mouse TRPC5, which involves conserved acidic amino acids in the putative outer pore region (Jung et al., 2003). 2-APB contains a positively charged group and clearly prevents the functional effect of Gd³⁺. Such an effect could be explained by competition at a common negatively charged acceptor site. However, the effects are opposite: lanthanides being stimulatory, 2-APB being inhibitory. Therefore, a more likely explanation is that they act at separate sites, with block of the channel by 2-APB indirectly occluding the activation by lanthanides. Further support for this hypothesis comes from the observation that 2-APB inhibits TRPC6, which does not contain the acidic amino-acid residues involved in stimulation of TRPC5 by lanthanides and is only inhibited by lanthanides.

At the concentrations employed in our study 2-APB is an inhibitor of IP₃Rs and modifies phospholipase C (PLC)γ-induced Ca²⁺ influx in TRPC3-transfected DT40 cells (Bootman et al., 2002; Ma et al., 2002). Therefore, our observation that there is no intracellular effect of 2-APB indicates IP₃Rs are nonessential for activation of TRPC5. A potential difficulty with using resistance to intracellular 2-APB as evidence for a lack of involvement of IP₃Rs is that block of IP₃Rs by 2-APB depends on the ambient IP₃ concentration – the required 2-APB IC₅₀ ranging from 10 μM to 1 mM (Bootman et al., 2002). However, this was not a problem in our experiments because IP₃ was not in the whole-cell patch pipette or in the solutions surrounding excised patches and, in most cases, TRPC5 was activated by Gd³⁺ rather than by a G-protein coupled receptor agonist that would stimulate PLC. Therefore, the ambient IP₃ concentration must have been low, making 2-APB potent as an inhibitor of IP₃ Rs. A study of mouse TRPC5 in Xenopus oocytes indicated a role for IP₃ and its receptor in TRPC5 activation (Kanki et al., 2001). In contrast, studies in HEK-293 and DT40 cells have found no effect of IP₃ and activation in the absence of IP₃R expression (Strubing et al., 2001; Venkatachalam et al., 2003). Our data on human TRPC5 are consistent with the conclusions of the latter studies.

Voltage-dependent block by an external positively charged substance that enters the ion pore can be explained by electrical repulsion. 2-APB contains positive charge and blocks TRPC5-mediated current and so the voltage-dependence we observe is consistent with 2-APB interacting with the ion pore of TRPC5, and thus amino acids presumably in and around the fifth and sixth membrane spanning regions. The voltage-dependence of the block is nevertheless complex and shows an intriguing inflexion in the same voltage range as the intrinsic IV of TRPC5. Therefore, it should not be excluded that 2-APB could, instead, produce voltage-dependent block because it is a gating modifier of TRP channels. Indeed, such an interpretation would more readily pave the way for a unifying hypothesis for 2-APB interaction with TRP channels, explaining why some are blocked, some unaffected and some stimulated. Sensitivity to 2-APB is an emerging common feature of TRP channels. At 10–100 μM, 2-APB is known to block nine of them: TRPC1, TRPC3, TRPC5, TRPC6, TRPV6, TRPM3, TRPM7, TRPM8 and TRPP2 (Ma et al., 2000; Voets et al., 2001; Delmas et al., 2002; Koulen et al., 2002; Schindl et al., 2002; Trebak et al., 2002; Lee et al., 2003b; Hanano et al., 2004; Hu et al., 2004; this study). High micromolar concentrations stimulate three of them: TRPV1, TRPV2 and TRPV3 (Chung et al., 2004; Hu et al., 2004). TRPC3, TRPV1 and TRPV6 have been observed to be unaffected by 2-APB under some conditions, while TRPV4 and TRPM2 are only known to be resistant (Schindl et al., 2002; Trebak et al., 2002; Chung et al., 2004; Hu et al., 2004; this study).

The expression level of TRP family members may affect the activation and pharmacological properties (Voets et al., 2001; Schindl et al., 2002; Trebak et al., 2002). TRPC3 and TRPV6 at low expression level show store-operated properties and both of them are sensitive to 2-APB, but they are resistant to2-APB when highly overexpressed. Regarding the experiments on TRPM2, we cannot exclude the possibility that resistance to 2-APB arises because of a high expression level. However, we find no evidence of such a phenomenon for TRPC5, which was always blocked by 2-APB whether expression levels were low or high, as indicated by ionic current amplitudes over almost two orders of magnitude.

A specific function of TRPC5 in neuronal growth cone formation has been proposed (Greka et al., 2003), but otherwise the physiological functions remain to be revealed. Our data now suggest an additional role in cell proliferation. 2-APB also inhibits native SOCs, consistent with the involvement of TRP proteins in this mechanism (Xu & Beech 2001; Flemming et al., 2003; Beech et al., 2004). The ‘CRAC' subtype of SOC is potentiated or inhibited by 2-APB (Prakriya & Lewis, 2001). Like TRP channels, store-operated Ca²⁺-entry has been associated with cell growth and proliferation (Hunton et al., 2002; Sweeney et al., 2002).

In conclusion, our data suggest a specific site for 2-APB on TRPC5 channels that is accessible only via an extracellular route and most probably involving interaction with the gating mechanism of the channel or a site within the electric field, resulting in an intriguing and previous unrecognised, albeit modest, voltage-dependence to the blockade. We observe differential sensitivity of TRP channels to 2-APB, but sensitivity is nevertheless a common feature, pointing to the existence of a conserved and functionally important drug-binding site on many members of this emerging family of channels.