20-hydroxyeicosatetraenoic acid (20-HETE) activates mouse TRPC6 channels expressed in HEK293 cells.

In the present study, we show that the eicosanoid compound, 20-hydroxyeicosatetraenoic acid (20-HETE), an important arachidonic acid metabolite, activates mouse TRPC6 in a stable, overexpressing HEK293 cell line, Hek-t6.11. Application of 20-HETE rapidly induced an inward, non-selective current in whole-cell recordings, which was inhibited by N-methyl-d-glucamine, 1.8 mm Ca2+, and 100 microM Gd3+ but remained unaffected by flufenamate and indomethacin. The current-voltage relationship obtained at low concentrations of 20-HETE (1-10 microM) demonstrated slight inward rectification, whereas the highest concentration of 20-HETE tested (30 microM) showed outward rectification, as shown previously for these channels using 100 microM 1-oleoyl-2-acetyl-sn-glycerol. Dose-response curves indicate that 20-HETE activated TRPC6 channels with an EC50 = 0.8 microM. Single channel analysis using inside-out patches revealed that 20-HETE increased open probability of mouse TRPC6 channels approximately 3-fold, and this was in a membrane-delimited fashion. Interestingly, 20-HETE did not provoke changes in intracellular Ca2+ concentrations. Thus, we have identified an arachidonic acid metabolite, 20-HETE, as a novel activator for a TRP family member, TRPC6.

cium release-activated calcium channel, a highly selective Ca 2ϩ ion channel first described in mast cells (7) and T lymphocytes (8), whereas non-selective cationic channels activated by store depletion have also been described (9). The molecular identity of store-operated channels has not yet been firmly established despite accumulating evidence that they could be members of the TRPC family (10 -12). The transient receptor potential (dTRP) channel, found in Drosophila eye, is a storeoperated Ca 2ϩ channel, whereas the trp-like (dTRPL) protein functions as a constitutively activated, non-selective cation channel (13)(14)(15). In mammals, seven dTRP homologs, TRPC1-TRPC7 (16 -18), have been found. Overexpression studies have demonstrated that these channels show important differences in a number of properties including their mode of activation, unitary conductance, and selectivity, which may or may not be a result of the expression system used (for reviews, see Refs. 4,17,and 19). Moreover, the formation of heteromeric TRPC channels has been reported, whose characteristics were quite distinct from those recorded for the individual, homomeric channels; such behavior was shown for coassembly of TRPC1/ TRPC3 (20) and TRPC1/TRPC5 (21). A number of studies have reported that some members of the TRPC family can be activated by products of the phospholipase C (PLC) signaling cascade, in a store depletion-independent manner. Indeed, diacylglycerol (DAG)-dependent activation has been shown for TRPC3/6/7 (18,22) as well as for TRPC1/TRPC3 heteromers (21). Moreover, products of phospholipase A2, arachidonic acid (AA) and its metabolites, and DAG lipase can also activate calcium-permeable channels (23)(24)(25). 20-hydroxyeicosatetraenoic acid (20-HETE) is the dominant arachidonic acid metabolite produced by cytochrome P-450 -hydroxylase enzymes (26). This bioactive eicosanoid compound is produced in various tissues, including, kidney (27,28), lung (26), and the vascular bed (28,29). Although its physiological importance is now emerging (26,30,31), the molecular mode of action of 20-HETE has yet to be established. It has been suggested that this compound represents an electropharmacological modulator capable of constricting various vascular and airway smooth muscle tissues by causing membrane depolarization and contraction (31,32). Since several recent studies have determined that vascular smooth muscle cells express TRPC6 channels (33,34), the present study was aimed at assessing the electrophysiological effects of 20-HETE on mTRPC6 channels in an overexpression system. The rationale was to test whether 20-HETE might activate a specific isoform of TRPC conductances in an attempt to explain its well characterized inotropic effects, which are not directly related to the activation of 3,4-dihydro-2H-pyran-sensitive Ca 2ϩ channels but rather to changes in non-selective membrane conductances (34,35). Our results conclusively demonstrate that application of 20-HETE activates mTRPC6 channels stably expressed in HEK293 cells and in smooth muscle cells.
Smooth Muscle Cell Culture-Male or female albino guinea pigs (Hartley, weighing 350 -450 g) were anesthetized by a lethal dose of pentobarbital sodium (50 mg/kg intraperitoneally) and sacrificed by abdominal exsanguinations. The trachea was excised aseptically and placed immediately on ice into sterile Krebs solution. The trachea was cut free of excess tissue and cut longitudinally on the opposite side of the smooth muscle. The epithelial cells were mechanically removed with a sterile cotton swab. The smooth muscle tissue was minced, washed in minimum essential medium (MEM) containing 200 M free Ca 2ϩ , and centrifuged at 80 ϫ g for 1 min. The pellet was resuspended and dissociated in 200 M Ca 2ϩ MEM with 640 units/ml collagenase (type IV), 10 units/ml elastase (type IV), and 20 g/ml DNase (type I), all from Sigma. The tissue was digested at 37°C for 3 ϫ 20 min. The cell suspension was then filtered through 100-m nylon cell strainer, and the filtrate was washed with 900 M Ca 2ϩ MEM. The cells were centrifuged for 10 min, and the pellet was resuspended in 1 ml of Opti-MEM supplemented by 2% fetal bovine serum and 1% penicillin-streptomycin. The cells were plated in 35 mm-dishes with ϳ10 3 cells for each dish, and after 30 min incubation at 37°C, the dishes were completed with 2 ml of Opti-MEM. Cells were used for electrophysiology experiments within 48 h.
Electrophysiology-Whole-cell currents in HEK293 and HEK-t6.11 cells were measured with the patch clamp technique (37) performed at room temperature using fire-polished patch pipettes (3-6 megaohms of uncompensated series resistance). Currents were recorded with an Axo-Patch amplifier (Axon Instruments, Union City, CA), controlled by homemade software. The standard holding potential was Ϫ40 mV, and membrane currents were filtered at 500 Hz and acquired at 1000 Hz. Single channel recordings were performed in the inside-out configuration with Sylgard-coated patch pipette. The membrane potential was set to Ϫ60 mV, and the signal was filtered at 1000 Hz and acquired at 2000 Hz.
In whole-cell configuration, the standard intracellular solution contained: 140 mM CsAsp, 1 mM CaCl 2 , 11 mM EGTA, 2 mM MgCl 2 , 18 mM NaCl, 10   Data Analysis-Analysis was performed with a homemade software written in Visual Basic. Channel activity is expressed as NP o , calculated for consecutive 5-s intervals. Data are expressed as mean Ϯ S.E.

RESULTS
Whole-cell experiments were performed on HEK-t6.11 cells stably expressing mTRPC6, as described previously (36). Following whole-cell break-in, an inward current was observed with a mean amplitude of Ϫ43 Ϯ 3.3 pA (n ϭ 31) at a membrane potential of Ϫ40 mV. Application of 1 M 20-HETE induced a rapid increase in current, as shown in Fig. 1A. This current remained activated in the continued presence of 20-HETE. When perfusion was switched to a 20-HETE-free solution, the level of current rapidly decreased back to control levels. No time-and dose-dependent desensitizations were observed. Fig. 1B illustrates the current to voltage relationship (I/V curve) of the 20-HETE-induced current. It is clearly evident that the 20-HETE-activated current displayed an I/V curve demonstrating a slight inward rectification. With Na ϩ and Cs ϩ as charge carriers, the reversal potential was found to be 3.4 Ϯ 1.5 mV (n ϭ 31), indicating a non-selective conductance.
To confirm that the 20-HETE-induced current was specific to mTRPC6 overexpression, experiments were also performed in parallel on non-transfected cells. The mean basal current level, in the absence of stimuli, was not different (Ϫ41 Ϯ 8 pA, n ϭ 6) from that recorded for HEK-t6.11 cells, but the current amplitude was not increased by the application of 1 M 20-HETE (Fig. 1C), and the I/V curves obtained in control and 20-HETE conditions were superimposable (Fig. 1D). The relationship between 20-HETE concentrations and current amplitude was then studied using the Hek-t6.11 cells. Current amplitudes induced by 20-HETE, at Ϫ40 mV, were measured as the difference between the current in the absence of stimuli and the current in the presence of the different concentrations of 20-HETE. The results presented in Fig. 1E show that the effects of 20-HETE are dose-dependent with an EC 50 value of 0.8 M. Fig. 2 (A and B) shows that for a concentration of 30 M 20-HETE, not only was the amplitude of the 20-HETE-induced current larger than at other doses, but more surprisingly, the I/V relation was different. Indeed, the curve was doubly rectifying, similar to what has been described for OAG-activated TRPC6 (22). To compare the OAG-and 20-HETE-induced currents in HEKt-6.11 cells, we performed experiments using two different concentrations of OAG, a membrane-permeable analog of DAG. When 100 M OAG was applied (Fig. 2C), the maximum amplitude of the induced current at Ϫ40 mV was Ϫ153 Ϯ 42 pA (n ϭ 5), and the I/V relation demonstrated double rectification (Fig. 2D). At a lower OAG concentration (50 M; Fig. 2E; Ϫ19.7 Ϯ 7 pA, n ϭ 4 out of 9), the I/V relation was similar to that obtained for 20-HETE concentrations lower or equal to 10 M, where both showed inward rectification (Fig.  2F). The reversal potential, using the same charge carriers, i.e. Na ϩ and Cs ϩ , was found to be near 0 mV under all conditions: 100 M OAG (mean ϭ 2.8 Ϯ 4.8 mV, n ϭ 5), 50 M OAG (mean ϭ 2.1 Ϯ 4.2, n ϭ 4), 30 M 20-HETE (mean ϭ 0.9 Ϯ 0.8, n ϭ 7). Similar results were obtained by applying carbamylcholine (1 M), which binds to endogenous muscarinic receptors and, through a G q protein, induces the production of inositol 1,4,5-triphosphate and DAG. The activated current was transient, and the I/V curve was similar (data not shown) to that recorded using high concentrations of OAG. No currents were triggered in wild-type HEK293 cells following application of either OAG or carbamylcholine (data not shown).
Experiments were then performed in isolated patches in the inside-out configuration to determine whether 20-HETE could activate mTRPC6 in a membrane-delimited fashion as already shown for DAG analogs (22). Application of 1 M 20-HETE to the bath induced a marked increase of the channel activity (Fig. 3A, n ϭ 7 out of 10). Fig. 3B shows the measured NP o over To better characterize the 20-HETE-induced current, the ionic composition of the external medium was modified. Inward currents were abolished when Na ϩ was replaced by NMDG ϩ as the only external cation charge carrier (Fig. 4A). Note that the basal current was also blocked in these conditions. The I/V curves with Na ϩ or NMDG ϩ are illustrated in Fig. 4B and reveal that the reversal potential was shifted to Ϫ25.6 Ϯ 0.5 mV (n ϭ 6) with NMDG ϩ in the bath.
The experiments performed above were conducted in a low Ca 2ϩ perfusion medium, and as shown in Fig. 4, C and D, perfusion of a 20-HETE solution containing 1.8 mM Ca 2ϩ decreased the amplitude of the current to basal levels, indicating that Ca 2ϩ blocked only the 20-HETE-induced current (n ϭ 4). In addition to Ca 2ϩ , several di-and trivalent cations can also block TRPC channels. Notably, Gd 3ϩ has been used to block capacitative currents at a concentration of 1 M, whereas 100 M blocks non-capacitative currents. Fig. 5A illustrates that, following current activation by perfusing a solution containing 1 M 20-HETE, the addition of 100 M Gd 3ϩ inhibited the 20-HETE-induced current (n ϭ 4), whereas perfusion of 1 M Gd 3ϩ was ineffective (data not shown) The non-specific cation channel blocker FFA was shown to enhance currents mediated by TRPC6, whereas currents mediated by TRPC3 and TRPC7 were inhibited (38). Contrary to this report, we found that the 20-HETE-triggered current in HEK-t6.11 cells was not sensitive to 100 M FFA (n ϭ 3, Fig. 5B). In human pulmonary arteries, the effects of 20-HETE are inhibited by preincubation with indomethacin, an inhibitor of cyclooxygenase 2 (COX-2), indicating that a metabolite of 20-HETE could be the active compound (39). We therefore tested whether this could be applied to our system. Fig. 5C demonstrates that preincubation of HEK-t6.11 cells with 1 M indomethacin for 10 min did not affect the 20-HETE-induced current (n ϭ 3).
In A7r5 smooth muscle cells, it has been reported that AVP activated a non-capacitative Ca 2ϩ entry directly regulated by arachidonic acid produced by the activities of PLC and DAG lipase (25). Similarly, dTRPL channels are activated by DAG and DAG analogs and by polyunsaturated fatty acids including arachidonic, linoleic, and linolenic acids (19,40). In HEK-t6.11cells, AA (10 M) activated a current with properties similar to those of the 20-HETE-induced current (Fig. 6, A and B,  n ϭ 4). Application of a non-metabolizable AA analog, ETYA, at a concentration of 10 M did not induce any currents (Fig. 6, C  and D, n ϭ 4). The concentration of ETYA used effectively blocks all pathways of AA metabolism including lipoxygenases, cyclooxygenase (COX), phospholipase A2, and cytochrome P-450, suggesting that the observed AA-induced currents are due to one of its metabolites.
Our next aim was to determine whether 20-HETE could induce a similar effect in a cell type that expresses TRPC6 endogenously. We therefore tested primary cultures of tracheal smooth muscle cells obtained from guinea pig where endogenous expression of TRPC6 was confirmed by reverse transcriptase-PCR (data not shown). In Fig. 7A, we show that 1 M 20-HETE induced an inward current in these cells with an average amplitude of Ϫ42 Ϯ 6 pA (n ϭ 3). Interestingly, removal of 20-HETE did not result in reversal of the induced current. The current-voltage relationship (Fig. 7B), however, is similar to that obtained with 20-HETE in Hek-t6.11 cells.

DISCUSSION
The data presented herein consistently and conclusively show that the eicosanoid compound, 20-HETE, activates mTRPC6 channels in Hek-t6.11 cells. TRPC6 channels belong to the subfamily TRPC3/TRPC6/TRPC7 (4), which displays the unique characteristic of being activated by DAG or DAG-permeant analogs. Application of OAG (100 M) activated a current demonstrating a characteristic doubly rectifying I/V curve in transfected cells (22), as well as in rabbit portal vein cells (38) and A7r5 smooth muscle cells (34), which express these channels endogenously. Similar results were obtained in HEK-t6.11 cells when OAG and carbamylcholine (data not shown) were applied. The I/V curve of the 20-HETE-activated current showed a slight inward rectification for low concentrations ranging from 1 to 10 M with a reversal potential near 0 mV, indicating the activation of a non-selective cation channel. At 30 M, the highest concentration used in our study, the I/V curve was similar to that obtained for 100 M OAG. Moreover, substitution of Na ϩ with NMDG ϩ abolished the inward current and shifted the reversal potential toward more negative values, in agreement with data obtained for OAG-activated TRPC6 currents (22). Other TRPC6-specific characteristics, such as blockage by 1.8 mM Ca 2ϩ and 100 M Gd 3ϩ , were also observed for the 20-HETE-activated current. FFA, a non-specific cation channel blocker, was shown to increase the current amplitude mediated by TRPC6 (34,38). However, we did not find that FFA had any effect on the 20-HETE-activated current.
Several reports have shown that TRPC members can form oligomeric complexes with subunits from the same subfamily. The resulting heteromeric channels were shown to behave differently than homomeric channels formed with either subunit alone. We verified that the overexpression of mTRPC6 did not result in the formation of such heteromeric channels by coimmunoprecipitation studies (data not shown). Moreover, the mRNA expression levels of the various TRPC subunits were not different between HEK293T and Hek-t6.11 cells (data not shown).
Previous reports have shown that polyunsaturated fatty acids, in particular AA, could induce non-capacitative currents in A7r5 smooth muscle cells and could activate the closely related dTRPL channels. In our model of overexpression, AA induced non-selective currents identical to those induced by 20-HETE. Moreover, lack of current activation when using a non-metabolizable form of AA (ETYA) supports the hypothesis that AA action appears to be mediated by a metabolite, most likely 20-HETE. Indeed, AA is metabolized primarily by cytochrome P-450 enzymes to produce EETs and 19-and 20-hydroxyeicosatetraenoic acids . It is now clear that these metabolites play a central role as second messengers in the regulation of renal vascular tone (27). The cytochrome P-450 4A family (CYP4A) catalyzes the formation of 20-HETE, all isoforms of which are expressed in human kidney (41). This raises the possibility that in HEK293 cells, AA could be metabolized into 20-HETE, and it would explain the similarity of the currents activated by AA and 20-HETE. We tested a second metabolite of AA, 14,15-EET, which is produced by cytochrome P-450 epoxygenases (30,41). This metabolite has been reported to activate large conductance Ca 2ϩactivated K ϩ channels and inhibit Cl Ϫ channels in vascular and airway smooth muscles (30,42). We found that 1 M 14,15-EET isomer did not trigger currents in our cell model (data not shown). Despite the molecular similarities between 20-HETE and 14,15-EET (both are hydrophobic and structurally related), these two molecules induce distinct responses. The mode of action of DAG analogs or polyunsaturated fatty acids has not yet been established. Based on results obtained with a PLC blocker (U73122), it has been proposed that 1-stearyl-2-acetyl-sn-glycerol, a membrane-permeant analog of DAG, and polyunsaturated fatty acids, when applied exogenously, stimulate PLC, which thus activates dTRPL channels indirectly (43). However, it has been shown that neither inositol 1,4,5-triphosphate nor phosphokinase C was able to activate TRPC6 (22). Similar phosphokinase C-independent activation of TRPC3/6/7 has been reported (for a review, see Ref. 19). This was further confirmed by the increase in channel activity in inside-out patches, outlining a direct effect on the channel. As we have found a similar effect for 20-HETE using isolated patches in an inside-out configuration, we can also conclude a direct effect of 20-HETE on mTRPC6.
Activation of mTRPC6 was reported to induce a cytosolic [Ca 2ϩ ] increase (36), and more specifically, DAG analogs induced an influx of Mn 2ϩ in hTRPC6-transfected cells (22). However, in HEK-t6.11 cells, 20-HETE (1-30 M) was not able to increase [Ca 2ϩ ] i or induce Mn 2ϩ influx (data not shown). One possibility could be related to the intrinsic nature of 20-HETE itself. Indeed, in Sf9 cells expressing the dTRPL channel, a significant correlation was found between intracellular Ca 2ϩ increase and current activation for linoleic acid and linolenic acid. However, 1-stearyl-2-acetyl-sn-glycerol and AA, strong activators of dTRPL current, were less potent or ineffective, respectively, in causing a rise in intracellular Ca 2ϩ (43). 20-HETE is a potent constrictor (EC 50 Ͻ 10 Ϫ8 M) of renal interlobular and afferent arterioles. It has been reported to inhibit the activity of Ca 2ϩ -activated K ϩ channels, which depolarize the membrane to open L-type voltage-dependent Ca 2ϩ channels (44,45). In this context, our results suggest that the current passing through mTRPC6 would drive the membrane potential toward 0 mV, the reversal potential of mTRPC6, and could thus favor the opening of voltage-dependent Ca 2ϩ channels. However, smooth muscle contraction induced by various agonists was only partially inhibited by L-type Ca 2ϩ channel blockers (46), indicating that L-type voltage-gated calcium channels do not entirely support the observed Ca 2ϩ influx. A role for non-selective cation channels in Ca 2ϩ entry during Ca 2ϩ wavelike oscillations in vascular smooth muscle has been proposed by Van Breemen and colleagues (for a review, see Ref. 47). The opening of non-selective cation channels triggered an influx of Na ϩ ions, which were then extruded by the Na ϩ /Ca 2ϩ exchanger, functioning in reverse mode, allowing Ca 2ϩ entry. These Ca 2ϩ oscillations were slowed by the addition of nifedipine but blocked by (2-aminoethoxy)-diphenylborane, a blocker of SOC (48) or 2,4-dichlorobenzamyl, a blocker of the Na ϩ /Ca 2ϩ exchanger.
Overall, the results herein describe the novel mechanism of activation of a TRP member by a bioactive eicosanoid, 20-HETE, of physiological relevance for a number of different tissues including kidney and vascular and airway smooth muscle. These data offer new insight into determining the components participating in the observed 20-HETE-regulated membrane depolarization of vascular smooth muscle cells.