TRPV2 Is a Component of Osmotically Sensitive Cation Channels in Murine Aortic Myocytes

Abstract— Changes in membrane tension resulting from membrane stretch represent one of the key elements in blood flow regulation in vascular smooth muscle. However, the molecular mechanisms involved in the regulation of membrane stretch remain unclear. In this study, we provide evidence that a vanilloid receptor (TRPV) homologue, TRPV2 is expressed in vascular smooth muscle cells, and demonstrate that it can be activated by membrane stretch. Cell swelling caused by hypotonic solutions activated a nonselective cation channel current (NSCC) and elevated intracellular Ca2+ ([Ca2+]i) in freshly isolated cells from mouse aorta. Both of these signals were blocked by ruthenium red, an effective blocker of TRPVs. The absence of external Ca2+ abolished this increase in [Ca2+]i caused by the hypotonic stimulation and reduced the activation of NSCC. Significant immunoreactivity to mouse TRPV2 protein was detected in single mouse aortic myocytes. Moreover, the expression of TRPV2 was found in mesenteric and basilar arterial myocytes. Treatment of mouse aorta with TRPV2 antisense oligonucleotides resulted in suppression of hypotonic stimulation-induced activation of NSCC and elevation of [Ca2+]i as well as marked inhibition of TRPV2 protein expression. In Chinese hamster ovary K1 (CHO) cells transfected with TRPV2 cDNA (TRPV2-CHO), application of membrane stretch through the recording pipette and hypotonic stimulation consistently activated single NSCC. Moreover, stretch of TRPV2-CHO cells cultured on an elastic silicon membrane significantly elevated [Ca2+]i. These results provide a strong basis for our purpose that endogenous TRPV2 in mouse vascular myocytes functions as a novel and important stretch sensor in vascular smooth muscles.

D etection of mechanical stimuli is essential for diverse biological functions including audition, touch, and maintenance of vascular myogenic tone. In the latter, elevation of intravascular pressure depolarizes vascular smooth muscle cells via membrane stretch. 1,2 This depolarization activates voltage-dependent L-type Ca 2ϩ channels (VDCC) and increases [Ca 2ϩ ] i , resulting in vasoconstriction and/or myogenic tone. 3 A large component of the elevation of [Ca 2ϩ ] i by myogenic tone can be inhibited by blockers of VDCC. However, some components are resistant to these agents, and instead can be accounted for by a separate nonselective cation channel that is permeable to Ca 2ϩ and also is activated by the intravascular pressure. 4,5 Because stretchactivated channels play obligatory roles in regulation of the myogenic tone, extensive studies have been performed to identify the molecular entity of these channels. Originally, a yeast MID1 gene product (MID1) was shown to be a eukaryotic stretch-activated channel, because CHO cells expressing MID1 responded to membrane stretch. 6 A member of the transient receptor potential channels (TRP), specifi-cally TRPC6, is sensitive to myogenic tone, possibly through the production of diacylglycerol derivatives (DAGs) by membrane stretch. 7 The vanilloid receptor (TRPV) family includes a cation selective channel. Two examples, TRPV1 and TRPV2, were first isolated from a cDNA library from rat sensory neurons. 8,9 The TRPV family has substantial sequence homology to many other membrane proteins including all members of TRPs and Caenorhabditis elegans (C elegans) OSM-9 channel. 10,11 Members of the TRPV are activated by a diverse range of stimuli such as heat, protons, lipids, and/or change in extracellular osmolarity. 11 Although TRPV1, a capsaicinsensitive TRPV in nociceptors, is known to integrate response to noxious stimuli, this channel subtype also may participate in normal bladder function by mediating mechanically evoked purinergic signaling in the urothelium. 12 TRPV4, a member of this family that is expressed in vascular endothelial cells as well as kidney, senses changes in cell volume and therefore is a sensor of osmotic pressure changes. 13,14 TRPV4 is also modulated by DAGs and heat stimulation. 15,16 More-over, OSM-9 channels are essential in C elegans neurons for several forms of sensory transduction, including osmo-and mechanosensation. 17 Although these findings suggest a fundamental relationship between TRPV and mechanosensors, it remains unclear whether TRPV can, in fact, function as stretch-activated channels, and their functional expression in vascular smooth muscles has not been demonstrated.
In the present experiments, we examined the possibility that TRPV2 (GRC/VRL1), a member of TRPVs, is involved in activation of NSCC by membrane stretch, in the setting of cell swelling in vascular smooth muscle cells. It is activated by heat and constitutively active in the presence of growth factors such as insulin growth factor I, platelet-derived growth factor, and serum. 18,19 By using immunocytochemistry and an antisense strategy, we first demonstrated that TRPV2 is expressed in mouse vascular smooth muscle cells, and then showed that it is activated by hypotonic stimulationinduced cell swelling. Moreover, we used recombinant expression of TRPV2 in CHO cells to demonstrate that this channel can be activated by membrane stretch as well as hypotonic stimulation and is responsible for elevation of [Ca 2ϩ ] i by cell stretch.

Materials and Methods
An expanded Materials and Methods section can be found in on online data supplement available at http://www.circresaha.org. Briefly, CHO cells were transfected with the recombinant plasmids pIRES-TRPV2 and pIRES-GFP. Single cells isolated from aorta and mesenteric and basilar artery were obtained using a dispersion procedure involving collagenase and papain. Electrophysiological experiments and data analysis were done as described elsewhere. 20 Stretch was applied to CHO cells using a chamber similar to that described by Naruse and Sokabe. 21 RT-PCR and Western Blotting analyses for TRPV expression were performed as described previously. 22 Oligodeoxynucleotides (ODNs) specific for mouse TRPV2 were designed and were introduced into intact aorta using a reversible permeabilization procedure. 23 Immunostained cells with antibodies were observed under a confocal laser scanning microscope.

Cell Swelling Induced by Hypotonic Stimulation Activates NSCC and Elevates [Ca 2؉ ] i in Mouse Aortic Myocytes
To demonstrate NSCC activated by hypotonic stimulationinduced cell swelling, we applied a hypotonic osmotic solution to mouse aortic myocytes while measuring membrane currents and [Ca 2ϩ ] i (Figure 1). These myocytes were voltage- Figure 1. Activation of inward currents and elevation of [Ca 2ϩ ] i by hypotonic stimulation-induced cell swelling in mouse aortic myocytes. A, Myocyte at a holding potential of Ϫ60 mV was superfused with mannitol compensated isotonic (mannitol sol., 310 mOsm) and then with 227 mOsm hypotonic solution in order to swell the cell. Ramp waveform command pulses were applied every 5 seconds (see the inset in B). Lines in the trace indicate current deflection elicited by these ramp pulses. B, I-V relationships of the inward currents induced by the hypotonic stimulation. I-V relationships numbered from 1 to 3 were obtained at the times indicated in A. C and D, Representative records describing effects of 227 mOsm hypotonic stimulation on Ca 2ϩ fluorescence ratio changes in mouse aortic myocytes without (C) and after application of pharmacological agents [D, pretreated with diltiazem (diltz) and caffeine (caff)]. E, Summary of Ca 2ϩ response of non-voltage-clamped mouse aortic myocytes to 227 mOsm hypotonic stimulation. Peak ratio change was obtained as ⌬ratio (peak) under each experimental condition. BHQ indicates tertbutylhydroquinone. Numbers in parentheses correspond to the number of cells studied. Probability value is a result of one-way ANOVA.
clamped at Ϫ60 mV, and ramp voltage waveforms were applied every 5 seconds. The application of the hypotonic solution (reduction in osmolarity from 310 to 227 mOsm) caused activation of inward currents in 79% mouse aortic myocytes (15 out of 19 cells, Figure 1A). Just after the activation of the inward currents, the reversal potential of the currents in the current-voltage (I-V) relationship was Ϫ6.8Ϯ1.3 mV (nϭ6), which was close to the theoretical equilibrium potential of monovalent cations under this experimental condition (ϷϪ9 mV, 1 and 2 arrows in Figure 1B). In 4 out of 6 cells, however, the I-V relationships were progressively shifted in the negative direction during this stimulation; the reversal potential was changed from Ϫ7.0Ϯ1.3 to Ϫ17.5Ϯ2.8 mV (nϭ4, PϽ0.05, 1 versus 3 arrows in Figure 1B). This shift of the I-V relationship is assumed to be caused by activation of Ca 2ϩ -activated Cl Ϫ currents (I Cl-Ca ) in mouse aortic myocytes, because removal of external Ca 2ϩ abolished the shift of the I-V relationship of the current and the change in the equilibrium potential of Clreversed the shift (see the online data supplement).
Effects of hypotonic stimulation-induced cell swelling on [Ca 2ϩ ] i in mouse aortic myocytes are illustrated in Figure 1C. In 56% of the cells studied (71 of 127 cells), 227 mOsm hypotonic stimulation induced elevation of [Ca 2ϩ ] i , which consisted of transient and sustained components. Neither Ca 2ϩ influx through VDCC nor release of Ca 2ϩ from Ca 2ϩ storage sites was responsible for the elevation of [Ca 2ϩ ] i , because pretreatment with 30 mol/L diltiazem (diltz), 10 mmol/L caffeine (caff), 10 mol/L tert-butylhydroquinone (BHQ), or the combination of these agents failed to affect the elevation of [Ca 2ϩ ] i . (Figures 1D and 1E).
To examine further this activation of NSCC and elevation of [Ca 2ϩ ] i due to hypotonic challenge, the osmotic stimuli were systematically changed, by superfusion with 310 (mannitol sol.), 278, 227 and 170 mOsm solution. Concentrations of cations and Cl Ϫ were kept constant during this protocol and equilibrium potential of monovalent cations, and Cl Ϫ was set at approximately Ϫ15 and ϩ15 mV, respectively. Inward currents were activated by stepwise changes in tonicity in the presence or absence of external Ca 2ϩ (Figures 2A and 2B). The I-V relationship of the inward currents revealed that only NSCC was activated in the absence of Ca 2ϩ (see the online data supplement). Fluorescence signals due to changes in [Ca 2ϩ ] i were measured in the same protocol ( Figures 2C and  2D). In the presence of Ca 2ϩ , the hypotonic stimulation elevated [Ca 2ϩ ] i , and the each response had both transient and sustained components ( Figures 2C and 2F). Both components were abolished in the absence of Ca 2ϩ ( Figures 2D and 2F). Figure 2E summarizes the current density of NSCC and I Cl-Ca which were estimated at ϩ15 and Ϫ15 mV, an equilibrium potential of Cl Ϫ and monovalent cations, respectively. I Cl-Ca was induced only in the presence of external Ca 2ϩ , suggesting that Ca 2ϩ entry through NSCC is required for the activation.

TRPV Modulators on Cell Swelling Induced-NSCC and Ca 2؉ Response
Because ruthenium red (RuR) is defined as an effective blocker of TRPVs, we examined effects of RuR on hypotonic stimulation-induced NSCC and elevation of [Ca 2ϩ ] i in mouse aortic myocytes ( Figure 3). As shown in Figure 3A, 10 mol/L RuR substantially inhibited inward currents activated by 227 mOsm solution, and the removal of RuR during the hypotonic stimulation resulted in recovery of these currents. Moreover, pretreatment of myocytes with 10 mol/L RuR effectively inhibited the 227 mOsm hypotonic stimulationinduced inward currents ( Figures 3B and 3Cb). In Figures 3D  and 3E, the same protocols were used to examine the effect of RuR on elevation of [Ca 2ϩ ] i by the hypotonic stimulation. The sustained component of the elevation of [Ca 2ϩ ] i was effec-tively inhibited by application of 10 mol/L RuR. In addition, the transient elevation was significantly inhibited by pretreatment with RuR. These inhibitory effects of RuR were not observed in any of the cells that were insensitive to the hypotonic stimulation (Figures 3Ca, 3D, and 3Fa).
TRPV4, which is activated by hypotonic stimulation and inhibited by RuR, is expressed in mouse aortic endothelial cells. 15,24 TRPV4 is also sensitive to DAGs such as phorbol didecanoate (PDD) and 4␣-PDD. 15,24,25 Effects of 4␣-PDD on [Ca 2ϩ ] i were examined in mouse aortic myocytes; these results are summarized in Figure 3G. Application of 1 mol/L 4␣-PDD elevated [Ca 2ϩ ] i in only 4 out of 31 myocytes; in contrast, all of the cells examined were sensitive to 227 mOsm hypotonic stimulation. Capsaicin can potently activate TRPV1. 8 Although in two myocytes there was slight increase of [Ca 2ϩ ] i by 1 mol/L capsaicin (caps), this response was much smaller than that to 227 mOsm hypotonic stimulation ( Figure 3G). These suggest that neither TRPV1 In panel Ca, data obtained from cells that were sensitive and insensitive to the hypotonic stimulation were pooled and averaged. Peak amplitude of NSCC was measured at Ϫ30 mV, where I Cl-Ca was negligible due to E Cl . D and E, Representative record of the effect of posttreatment and pretreatment with 10 mol/L RuR, respectively, on 227 mOsm hypotonic stimulation-induced change of Ca 2ϩ fluorescence ratio. Trace indicated by "a" denotes the Ca 2ϩ signal of a cell insensitive to 227 mOsm hypotonic stimulation. Fa and Fb, Summary of effects of 10 mol/L RuR on 227 mOsm hypotonic stimulation-induced sustained and peak change of Ca 2ϩ fluorescence ratio, respectively. G, Negative effects of 1 mol/L 4␣-PDD and 1 mol/L capsaicin (caps) on peak Ca 2ϩ fluorescence ratio in mouse aortic myocytes. *PϽ0.05 and **PϽ0.01, statistically significant difference.
nor TRPV4 is involved in the NSCC in mouse aortic myocytes, which is activated by cell swelling through hypotonic stimulation and can be blocked by RuR.

TRPV2 Transcript and Protein Expression in Mouse Vascular Myocytes
In Figure 4A, we evaluated the expression of other TRPVs mRNA in an effort to find a candidate of NSCC activated by hypotonic stimulation-induced cell swelling. Total RNA isolated from acutely dispersed mouse aortic, mesenteric, and basilar arterial myocytes was subjected to RT-PCR. In mouse aortic myocytes, the transcripts of mRNA of TRPV2 and TRPV4 were present, whereas heat-and RuR-sensitive TRPV3 was not expressed (nϭ5 to 6). Expression of TRPV2 was also found in both mesenteric and basilar arterial myocytes (nϭ5 to 6). Immunocytochemistry using anti-TRPV2 antisera demonstrated that TRPV2 proteins were expressed in acutely dissociated mouse aortic myocytes ( Figure 4B), whereas no immunoreactivity was detected in myocytes treated with Alexa488-conjugated secondary antibody alone (data not shown) or with preabsorption of anti-TRPV2 antibody by the immunizing peptide (IP, Figure 4B). Moreover, in mesenteric and basilar arterial myocytes, which were positively stained with anti-smooth muscle-specific ␣-actin antibody, expression of TRPV2 protein was confirmed by the immunocytochemical analysis.

TRPV2 Functions as Ca 2؉ Entry Channel Activated by Cell Swelling in Mouse Aortic Myocytes
To determine whether the endogenously expressed TRPV2 protein functions as NSCC and contributes to elevation of [Ca 2ϩ ] i by hypotonic stimulation-induced cell swelling, we next utilized an antisense ODN designed to be specific for TRPV2. Mouse aortic strips were organ-cultured for 4 to 5 days after the treatment with sense and antisense ODNs. The expression of TRPV2 protein was significantly decreased in myocytes isolated from strips treated with the antisense ODN, whereas substantial TRPV2 protein remained in sense ODN-treated preparations ( Figures 5A  and 5B). Correspondingly, the density of NSCC activated by 227 mOsm hypotonic stimulation was markedly decreased with the antisense ODN compared with cells treated with the sense ODN ( Figures 5C and 5E). We also examined the ability of TRPV2 to increase [Ca 2ϩ ] i after hypotonic stimulation (Figures 5D and 5F). Treatment with the antisense ODN significantly decreased the elevation of [Ca 2ϩ ] i in response to 227 mOsm hypotonic stimulation. As a control, we tested the effects of the ODNs on ATP-induced membrane currents and elevation of [Ca 2ϩ ] i and observed no differences in cells from senseversus antisense-ODN-treated aorta ( Figures 5E and 5F).

Transient Expression of TRPV2 in CHO Cells
Based on these findings, the characteristics of NSCC activated by hypotonic stimulation in mouse aortic myocytes were compared with those due to the activation of mouse TRPV2, which was transiently expressed in CHO cells. As shown in Figure 6A, membrane currents recorded in CHO cells transfected with TRPV2 were strongly inhibited by 10 mol/L RuR and this RuR-sensitive current reversed at Ϫ7.2Ϯ1.9 mV (nϭ4). In contrast, 10 mol/L RuR had no effect on membrane currents recorded from native CHO cells (control-CHO, Figure 6C). Noisy channel events were consistently recorded when outsideout patches were obtained from TRPV2-CHO cells ( Figure  6B, nϭ7). Application of 10 mol/L RuR inhibited these events. This RuR sensitive single channel current reversed at Ϫ3.0Ϯ1.1 mV (nϭ4) when 140 mmol/L K ϩ was included in the pipette solution. The replacement of 110 mmol/L Cl Ϫ with equimolar aspartate Ϫ and 70 mmol/L K ϩ with Cs ϩ did not change the reversal potential of this channel current (Ϫ5.7Ϯ1.7 mV, nϭ3).

Activation of TRPV2 by Membrane Stretch and Hypotonic Stimulation
We next examined whether TRPV2 can be activated by membrane stretch and hypotonic stimulation at the single channel level. As shown in Figure 7A, application of a negative pressure of 15 cm H 2 O to the recording pipette did not elicit any channel activity, whereas that of 30 cm H 2 O markedly activated channel currents. At Ϫ40 mV, reapplication of the negative pressure of 30 cm H 2 O elicited channel currents that were inward. Even after the cessation of the membrane stretch, sporadic channel openings were still observed. Among GFP-positive TRPV2-CHO cells, 42% cells responded ( Figure 7C, averaged current amplitude: 1.73Ϯ0.47 pA at ϩ30 mV; PϽ0.05 versus control). This pattern of channel openings was never observed in patches from control-CHO cells (Figures 7B and 7C, 7.7%, 0.08Ϯ0.08 pA at ϩ30 mV). In Figure 7D, a TRPV2-CHO cell was superfused with 227 mOsm hypotonic solution under the cell-attached patch-clamp configuration. The hypotonic stimulation activated channel currents in TRPV2-CHO cells, although not in control-CHO cells ( Figure 7E).

Elevation of [Ca 2؉ ] i in TRPV2-CHO Cells by Cell Stretch
Activation of TRPV2 in mouse aortic myocytes caused elevation of [Ca 2ϩ ] i . To confirm that sarcolemmal stretch contributes to elevation of [Ca 2ϩ ] i through TRPV2, experimental protocols as shown in the schematic diagram in Figure  8 were performed in TRPV2-and control-CHO cells. Cells were cultured in a chamber in which the bottom was covered with a thin elastic silicon membrane. Deflection of this membrane caused adherent cells to stretch by approximately 18% (see the online data supplement), ie, elongation of cells elicited elevation of [Ca 2ϩ ] i in both TRPV2-and control-CHO cells (Figures 8A and 8B). Both the number of cells responding and the change in fluorescence signal ( Figure 8C, 61.9%, ⌬ratio: 0.052Ϯ0.009; PϽ0.05) were significantly greater in TRPV2-CHO cells than those in the control (33.3%, ⌬ratio: 0.016Ϯ0.004). These findings suggest that cells expressing TRPV2 are more susceptible to the cell stretch than controls.

Discussion
Our study demonstrates that TRPV2, one of TRPVs, is expressed in mouse aortic myocytes and acts as an essential molecular component for nonselective cation channel (NSCC) activated by hypotonic stimulation-induced cell swelling, and shows that this occurs with elevation of [Ca 2ϩ ] i . This conclusion, that TRPV2 is an essential component of the hypotonic response of NSCC in murine vascular myocytes is based on the following lines of evidence. (1) Activation of NSCC and the elevation of [Ca 2ϩ ] i , which were induced by hypotonic solutions, are both due to cation influx and are inhibited by RuR. (2) TRPV2-specific immunoreactivity was detected in myocytes from mouse mesenteric, basilar as well as aortic myocytes. (3) In mouse aortic myocytes, application of TRPV2 antisense ODN, but not that of sense ODN, substantially decreased TRPV2 protein expression and markedly reduced both the activation of NSCC and the elevation of [Ca 2ϩ ] i by hypotonic stimulation.
In the present study, single channel recordings established that TRPV2 is activated by hypotonic stimulation. Furthermore, TRPV2 is sensitive to membrane stretch induced by negative pressure through patch pipettes, and this involves a marked elevation of [Ca 2ϩ ] i . These results indicate that membrane stretch is a plausible activator of TRPV2. Although TRPV2 is activated by high temperature (Ͼ52°C) and growth factors, the present study provides the novel evidence that TRPV2 functions as a mechanosensor in various organs including vascular smooth muscles. CHO cells expressing MID1 responded to similar membrane stretch to that in the present study, 6 showing that MID1 is an eukaryotic stretchactivated channel. In addition, TRPC6 can be activated by change of membrane tension via myogenic tone in rat cerebral artery. 7 TRPV4 might also have a certain role in detecting mechanostimulation because TRPV4 is activated by hypotonic stimulation. However, mechanisms involved in opening of these channels by membrane stretch have not been defined. TRPVs, including TRPV2, have ankyrin repeats in the N-terminal region. Deletion of the ankyrin repeats abolished heat-activation of TRPV1 26 and TRPV4. 27 Because the ankyrin repeats interact with certain cytoskeletal proteins, this region of TRPV2 might be also important for acceptance of applied mechanical signals.
RT-PCR, immunocytochemical, and Western analyses revealed that mouse aortic myocytes express TRPV2. More interestingly and importantly, TRPV2 is present in basilar as well as mesenteric arterial myocytes. In rat cerebral artery, 7 treatment with the antisense ODN specific for TRPC6 suppressed pressure-induced myogenic tone and depolarization, and abolished hypotonic stimulation-induced NSCC. Mouse vascular smooth muscles express TRPC1, TRPC3, TRPC4, and/or TRPC6. 28,29 Although expression of TRPV2 and TRPC6 in the resistance arteries has not been quantitatively compared, the present finding demonstrates that TRPV2 may play a certain role in regulating vascular tones in these peripheral arteries. TRP channels can be activated/modulated by stimuli such as G protein-coupled receptor activation, store depletion, activated G protein, DAGs, inositol trisphos-phate, and Ca 2ϩ . 9,30 -32 In principle, a number of different TRPs could be also sensitive to membrane stretch and/or hypotonic stimulation. Although our results clearly demonstrate that TRPV2 in vascular myocytes is one of essential component of NSCC activated by membrane stretch and hypotonic stimulation-induced cell swelling, experiments using intact arteries will develop total understanding of the physiological significance of TRPV2 for smooth muscle regulation, in particular, in resistance arteries.
Even though TRPC6 is sensitive to myogenic tone, this may arise from DAGs produced by activation of phospholipase C, after cell swelling. Thus, changes in myogenic tone can play an obligatory role in the activation of TRPC6. 7 Moreover, activation of TRPV4 by hypotonic stimulation is mediated by phosphorylation of the tyrosine residue by Src family tyrosine kinase, 33 suggesting that hypotonic stimulation releases a ligand that activates this kinase. Although the tyrosine residue is not conserved in TRPV2 33 and stimulation of the kinase by hypotonic stimulation has not been confirmed in mouse aorta, the possibility that cell swelling produces an endogenous ligand that activates TRPV2 cannot be ruled out. Local change of membrane tension produced by membrane stretch could affect channels, receptors, lipids, and/or enzymes, and the locally produced ligand, in particular, a certain lipid, may modulate the neighboring channels. Activity of TRPV1 and TRPM7 is regulated by inositol lipids that are closely localized with these channels. 34,35 Taken together, the mechanism underlying the activation of TRPV2 by membrane stretch is remained to be determined.
RT-PCR analysis revealed that mouse aortic myocytes express TRPV4 as well as TRPV2, but not TRPV3. However, 4␣-PDD, 15,24,25 an activator of TRPV4, did not affect [Ca 2ϩ ] i in mouse aortic myocytes. Thus, TRPV4 appears not to serve as a major component of NSCC activated by hypotonic stimulation and is unlikely to contribute to the elevation of [Ca 2ϩ ] i in mouse aorta under the present experimental conditions. Definition of the functional roles of TRPV4 in vascular smooth muscles is required for understanding of TRPVs in vascular beds where it is expressed.
In conclusion, we have shown that TRPV2 is a major component of native Ca 2ϩ -permeable cation channels, which respond to membrane stretch in mouse aortic myocytes, and that the activation of TRPV2 is responsible for mechanosensitive membrane depolarization.