Regulation of TRPV4 by the Extracellular Osmolarity

A number of cues indicated to investigators who first studied TRPV4 that it may be an osmotically or mechanically sensitive channel. These included the knowledge that OSM-9 from C. elegans is involved in behavioral responses to increases in extracellular osmolarity and mechanical stimulation in the worm [22]. Furthermore, some of the tissues where mammalian TRPV4 expression is found are exposed to changes in osmolarity (the kidney) or are involved in osmoregulation (the circumventricular organs) or mechanosensation (hair cells in the ear). Changes in osmolarity provide both an osmotic and mechanical stimulus, owing to cell swelling and membrane stretching.

TRPV4 is highly sensitive to changes in extracellular osmolarity around the normal osmolarity of body fluids. Reductions in the extracellular osmolarity result in increases in [Ca²⁺]_i and membrane currents (Fig. 9.1A,B), whereas osmolarities above 300 mosmol l⁻¹ decrease both [Ca²⁺]_i and currents (Fig. 9.1C) in those cells displaying spontaneous activity [11]. Significant changes in Ca²⁺ influx were observed in response to changes in extracellular osmolarity of as little as 1 percent [12]. Stronger reductions in osmolarity result in larger increases in [Ca²⁺]_i and currents, which are often not graded but show a slow increase followed by a very rapid response to a high level [11]. This behavior probably reflects the positive feedback effect of Ca²⁺ entry on the channel (see below). In addition to relatively stable increases in [Ca²⁺]_i Liedtke et al. also reported that some cells respond to decreases in osmolarity with Ca²⁺ oscillations [12].

Responses to reduced osmolarity show differences depending on the expression system, recording conditions, and laboratory. Gao et al. [23] reported no elevation of basal Ca²⁺ in TRPV4-expressing cells and found very poor Ca²⁺ responses to hypotonic solutions at room temperature. Responses were nearly fourfold larger at 37°C. Others have observed strong [Ca²⁺]_i responses to reduced osmolarity at room temperature [11,24]. A more intact intracellular environment is probably important for the response to hypotonicity. Current responses to hypotonic solutions are slow and weak in ruptured patch whole-cell recordings compared to those in perforated-patch recordings. In ruptured patch-clamp recordings, current responses to hypotonic solutions were also much smaller (ninefold) than responses to 4α-phorbol esters, potent ligands that activate TRPV4 [25]. Other groups report no response of TRPV4 to reduced osmolarity in CHO cells, but responses in HEK cells [26]. Other factors that may influence the response to changes in osmolarity include proteins like microtubule-associated protein 7 (MAP7) that interact with TRPV4 [27].

Consistent with the data on the osmosensitivity of heterologously expressed TRPV4, TRPV4^−/− mice displayed defects in osmoregulation [28,29]. TRPV4 has also been suggested to be involved in the nociceptive response of primary sensory neurons [18,30], and the [Ca²⁺]_i response of airway smooth muscle cells to hypotonic stimulation [16]. In endothelial cells, part of the response to hypotonic solutions is abolished in TRPV4^−/− mice [31]. In C. elegans lacking the TRPV isoform OSM-9, TRPV4 targeted to the appropriate sensory neurons restored responses to osmotic stimuli [32]. Surprisingly, in this context, TRPV4 rescues responses to hypertonic solutions, whereas in mammalian cells, channel activation is seen on application of hypotonic solutions. The reason for this difference is unclear. Perhaps worm neurons, in contrast to mammalian cells, can generate an appropriate messenger to activate TRPV4 in response to hypertonic stimuli.

Mechanosensitivity of TRPV4

The involvement of TRPV4 in responses to cell swelling and its localization in tissues like the inner ear, sensory neurons, and endothelial cells that are known to express mechanosensitive channels raise the question of whether TRPV4 is a mechanosensitive channel. We saw no activation of TRPV4 in response to membrane stretch by applying suction to the patch pipette in the cell-attached mode in HEK293 cells [11]. A more recent study has, however, shown that TRPV4-expressing cells respond in the whole-cell mode to cell inflation with a current activation [26].

TRPV4’s responsiveness to noxious mechanical stimuli has also been suggested from a study of TRPV4^−/− mice [26,28]. TRPV4 also rescues mechanosensitivity in sensory neurons from C. elegans lacking OSM-9 [32]. TRPV4 was a major candidate for a mechanosensitive channel in hair cells in the cochlea involved in hearing. TRPV4^−/− mice displayed a delayed onset hearing loss and an increased susceptibility to acoustic injury [33]. The precise role of TRPV4 in hearing has not been identified, but the channel does not appear to be the mechanosensitive channel in hair cells.

The expression of TRPV4 in tissues that show flow- or shear stress-dependent Ca²⁺ entry like flow-sensitive nephron segments and the vascular endothelium make TRPV4 a candidate for a channel sensitive to these stimuli. Increases in [Ca²⁺]_i sensitive to ruthenium red, an inhibitor of TRPV channels (see below), have been reported in response to shear stress in HEK293 cells expressing TRPV4 at 37°C, but not at room temperature [23]. To date, there has been no demonstration of TRPV4 involvement in responses to these stimuli in native tissues. An involvement in flow-sensitive Ca²⁺ increases in the nephron is not consistent with a report of a mainly basolateral localization in tubule cells [21]. Surprisingly, increases in viscosity produced by adding dextran to the extracellular solution have been reported to increase currents in Hela cells expressing TRPV4 [20].

Mechanism of Activation by Hypotonic Solutions and Mechanical Stimuli

In the original demonstrations that TRPV4 could be regulated by changes in the extracellular osmolarity, it was unclear how the channel responds to these changes. We were careful to point out that the channel may respond directly to changes in osmolarity or may be activated by an endogenous signaling cascade sensitive to swelling [11]. A delay of at least twenty seconds before TRPV4-expressing cells respond to osmotic stimuli, however, suggests that a slower process (e.g., cell swelling or signaling pathways activated by cell swelling) is involved in the response. Because many cellular signaling cascades have been demonstrated to be activated by changes in cell volume, there is no lack of prospective candidates [34]. In contrast to volume-regulated anion channels, which are activated by a mechanism involving a swelling-induced reduction in intracellular ionic strength, and tyrosine kinase activation via small GTP-binding proteins, TRPV4 did not respond to infusion of GTPγS or to reductions in intracellular ionic strength [35]. Nonreceptor tyrosine kinases have, however, been implicated in the response of TRPV4 to hypotonic solutions, and hypotonic stress has been reported to increase tyrosine phosphorylation of native and heterologously expressed TRPV4 by members of the Src family of tyrosine kinases [36]. When coexpressed with TRPV4, Lyn, a tyrosine kinase of this family, dramatically increased phosphorylation of Tyr253 (Figure 9.2) in response to treatment with hypotonic solutions [36]. Mutation of this tyrosine resulted in a loss of the Ca²⁺ response of TRPV4-expressing cells to hypotonic solutions, leading to the suggestion that the response of TRPV4 to hypotonic cell swelling is mediated by tyrosine kinase-dependent phosphorylation. However, neither Vriens et al. [24] nor we (unpublished data) could confirm the loss of responsiveness to hypotonic solutions with this mutant.

Watanabe et al. showed that TRPV4 can be activated by products of arachidonic acid breakdown [37]. Arachidonic acid and endogenous activators of TRPV4, like the cannabinoid anadamide, that yield arachidonic acid are metabolized in a cytochrome P450 epoxygenase-dependent manner to epoxyeicosatrienoic acids (EETs). Inhibition of the cytochrome P450 epoxygenase strongly inhibited channel activation by arachidonic acid, and application of EETs activated heterologously expressed and endothelial TRPV4 [31,37]. Responses to hypotonic solutions were reduced by structurally unrelated inhibitors of PLA₂ and by inhibitors of CYP450 epoxygenase [24,31] and were increased by upregulation of the CYP450 epoxygenase [31]. From these data, it is probable that the sensitivity to changes in extracellular osmolarity that TRPV4 confers on cells results from activation of an endogenous signaling cascade activated by changes in cell volume: swelling-induced activation of PLA₂, release of arachidonic acid from membrane phospholipids, followed by CYP450 epoxygenase-dependent metabolism of arachidonic acid to EETs, which then activate the channel. There is also evidence that, like the effects of hypotonic solutions, the activation of TRPV4 by increased extracellular viscosity is also PLA₂ dependent [20]. However, it remains unclear whether EETs activate TRPV4 by a direct interaction with the channel or whether other mediators are involved.

A further interesting open question is how basal channel activity is generated and how reductions in [Ca²⁺]_i and currents occur in response to hypertonic solutions. A simple explanation would be that PLA₂ is active under isotonic conditions and that activity can be decreased by hypertonic solutions. However, Vriens et al. found that inhibitors of PLA₂ have no effect on the basal activity of TRPV4 [24]. Thus, different mechanisms are likely to be involved in the generation of basal TRPV4 activity and in the response of the channel to hypotonic solutions.

Activation of TRPV4 by Phorbol Ester Derivatives

A very important finding for TRPV4 was the identification of phorbol ester derivatives as channel activators (Figure 9.1D) [25]. TRPV4 was found to be activated by 4α-phorbol ester derivatives with EC₅₀ values around 0.2 μM for 4α-phorbol dide-canoate (4αPDD). These 4α-phorbol ester derivatives are commonly used as negative controls for the 4β-phorbol esters. The latter are widely used as activators of protein kinase C, (PKC) but also affect other proteins with C1-domains independently of PKC. The 4α-phorbol ester derivatives are not known to activate other ion channels and therefore represent specific tools to activate TRPV4 and study TRPV4-mediated effects. Under the conditions used in the study of Watanabe et al., the 4β-phorbol ester phorbol myristate acetate (PMA) was a less potent activator of TRPV4 than the 4α derivate, and acted as a partial agonist [25]. In ruptured patch whole-cell experiments, 4α-phorbol esters are more effective than hypotonic solutions in activating TRPV4 [25], but in more intact cells, [Ca²⁺]_i increases in response to hypotonic solutions and 4αPDD are similar [24]. The effects of 4α-phorbol ester derivatives are not mediated by the PLA₂ pathway that is involved in activating TRPV4 by hypotonic solutions, because 4α-phorbol ester-mediated activation still occurs in the presence of PLA₂ inhibitors [24]. However, although the effects of hypotonic solutions and phorbol esters are mediated by different pathways, the effects of heat (see below) and 4αPDD seem to share a common mechanism. By analogy to TRPV1, for which a YS motif in the S2–S3 linker was shown to be involved in the binding of capsaicin [38], Vriens et al. identified a YS motif at the N-terminal end of S3 (Tyr555 and Ser556, Figure 9.2) [24]. Mutations of Tyr555 led to a loss of responsiveness to 4αPDD, but not to arachidonic acid, lending further support to the hypothesis that different pathways are involved in the responses to cell swelling and phorbol esters. Furthermore, the Tyr555 mutation also led to a loss of the response to heat (see below). The independence of pathways for the response of TRPV4 to cell swelling and arachidonic acid metabolism, and the responses to 4α-phorbol esters and heat have recently been confirmed in endothelial cells [31].

Studies performed at 37°C rather than at room temperature report different effects of 4β-phorbol ester derivatives [23,39]. At the higher temperatures, 4β-phorbol ester derivatives had larger effects than at room temperature, and a large part of these effects was prevented by PKC inhibitors. These results suggest that the 4β-phorbol ester derivatives may have PKC-dependent and PKC-independent effects on TRPV4.

Regulation of TRPV4 by Temperature

Members of the TRPV family, with the exception of TRPV5 and TRPV6, form temperature-sensitive channels with temperature sensitivities ranging from warm (TRPV3/TRPV4) through hot (TRPV1), to very hot (TRPV2) [40]. TRPV4 has a temperature activation threshold between 25°C and 33°C in HEK293 cells [19,41], and around 27°C in Xenopus oocytes [19]. Responses to warming show desensitization, but this is incomplete, leaving a spontaneously active current component at temperatures around 37°C. TRPV4 responses to heat are also observed in cell-attached patches, but not in cell-free patches [41], suggesting that cellular components not present or functional in excised patches are involved in the production of a messenger responsible for the heat response. Heat-activated currents are smaller than those in response to 4αPDD. The current density in response to warming to 38°C was about one fifth of that in response to 4αPDD. However, there seem to be similarities in the responses to heat and 4αPDD because both responses are lost in the mutants of the YS motif in S3 [24]. This suggests that heat may influence the production of an as-yet-unidentified endogenous messenger that binds to a region of TRPV4 involving S3.

Responses of TRPV4 to warm temperatures and its expression in sensory neurons, keratinocytes, and in the hypothalamus indicate a role for TRPV4 in thermosensation and thermoregulation. TRPV4^−/− mice show decreased frequencies of nerve discharge in response to thermal stimulation, a decrease in the number of responsive fibers, and longer latencies to escape from thermal stimuli in a model of thermal hyperalgesia [42]. A lack of TRPV4 has no effect on escape latency in response to heat in the absence of hyperalgesia [26,28,42], with the exception of a study that described an increase in tail withdrawal latency to moderately hot temperatures [43]. The latter study also showed that TRPV4^−/− mice showed a preference for warmer floor temperatures. TRPV4 expression in the hypothalamus is an indicator of a possible role of TRPV4 in thermoregulation. However, Liedtke and Friedman found no difference in body temperature or differences in the temperature response to cold stress in TRPV4^−/− mice [28].

Ca²⁺ Dependence of TRPV4

Most Ca²⁺-permeable channels show some feedback regulation by Ca²⁺ to prevent deleteriously large increases in [Ca²⁺]_i or to shape the time course of channel activity. There is clear experimental evidence that, in addition to allowing Ca²⁺ to permeate, TRPV4 is closely regulated by [Ca²⁺]_i. As described above, responses of TRPV4 to hypotonic solutions, phorbol esters, or heat are transient with rapid activation followed by inactivation/decay (Figures 9.1B, D). Ca²⁺ is involved both in the decaying phase and the activation phase of TRPV4; in the latter, it is not only a charge carrier, but also a potentiator of channel activity.

Outward currents mediated by TRPV4 are strongly reduced by removing extracellular Ca²⁺. Similarly, removing extracellular Ca²⁺ abolishes spontaneous TRPV4 currents in parallel with the reduction in [Ca²⁺]_i [44]. There is a very close correlation between [Ca²⁺]_i and the amplitude of TRPV4 currents. Replacement of extracellular Ca²⁺ by Ba²⁺ or Sr²⁺, which also permeate well (see below), leads to a reduction in spontaneous inward and outward currents, a result that indicates that these ions are unable to substitute for Ca²⁺ in the mechanism involved in generating spontaneous channel activity. In the absence of extracellular Ca²⁺, or when Ca²⁺ is replaced by Ba²⁺, activation of TRPV4 by hypotonic solutions or 4αPMA is slow, taking several minutes rather than seconds to reach a maximum.

Ca²⁺ could either regulate the activity of TRPV4 by influencing Ca²⁺-dependent steps in signaling pathways leading to channel activation, or by binding to the channel or to channel-associated proteins. TRPV4 does not have an obvious Ca²⁺-binding site like an EF hand, suggesting that Ca²⁺ does not bind directly to the channel. Many other TRP channels have calmodulin (CaM)-binding sites in their C-termini, but regulation via Ca²⁺ and CaM has only been convincingly demonstrated for a few isoforms, including the TRPV isoforms TRPV1 [45] and TRPV6 [46,47]. We identified a CaM-binding site similar to that in the C-terminus of human TRPV6 [46] in the intracellular C-terminus of TRPV4 (Figure 9.2) [44]. This binding site is a highly conserved helical stretch of amino acids, VGRLRRDRWSSVVPRV, starting at position 814.

In in vitro CaM-binding experiments, mutations of amino acids within this region, including the point mutation W822A, prevented Ca²⁺-dependent CaM binding [44]. For the C-terminal CaM-binding site of TRPV6, similar mutations that prevent Ca²⁺-dependent CaM binding decrease the rate of a slower phase of channel inactivation [46]. In strong contrast to TRPV6, we found that the mutations that prevent CaM binding to C-terminal peptides of TRPV4 influence Ca²⁺-dependent potentiation and not inactivation of the channel.

Wild-type TRPV4 shows strong potentiation on the addition of Ca²⁺ following channel activation by hypotonic solutions or 4α-phorbol esters in a Ca²⁺-free medium [44]. After transient expression, the full-length channels with mutations that prevent Ca²⁺-CaM binding showed no potentiation of current responses to 4αPMA on Ca²⁺ addition, and only slow activation by 4αPMA in the presence of extracellular Ca²⁺. The speed of activation in Ca²⁺ resembled that of TRPV4-wt in the absence of Ca²⁺ or when Ba²⁺ replaced Ca²⁺.

From these data, we suggest that there is a Ca²⁺-dependent component of TRPV4 activation. Irrespective of the primary activator, hypotonic medium, lipophilic ligands, or heat, Ca²⁺ entering through TRPV4 binds to CaM, and Ca²⁺-CaM interacts with the C-terminal binding domain, resulting in positive feedback activation of TRPV4, increasing both the speed and amplitude of the response. It remains unclear, however, how Ca²⁺-CaM binding exerts its effects on channel activity and what the physiological role of this mechanism may be.

If unchecked, positive feedback involving Ca²⁺ could lead to large, potentially harmful increases in [Ca²⁺]_i and, thus, some kind of inhibitory mechanism is necessary to turn the channel off. Ca²⁺ limits the entry of Ca²⁺ through other Ca²⁺-permeable channels by feedback inhibition. Likewise, Ca²⁺ is involved in the decay of TRPV4 currents and in controlling channel availability. Large, fast TRPV4 current responses decay rapidly to levels below those of spontaneous currents before activation (see, e.g., Figure 9.1B), even in the continuous presence of the activating stimulus. Current responses in the absence of Ca²⁺ are in general smaller, and display slower activation kinetics (as described above) but also much weaker current decay, even when current amplitudes similar to those in the presence of Ca²⁺ are attained. With increasing extracellular [Ca²⁺], current responses to 4α-phorbol esters activate more rapidly (as expected from the Ca²⁺-dependent potentiation described above), become smaller in amplitude, and decay more completely and rapidly [48]. Watanabe et al. [48] also saw a reduction in current amplitude as [Ca²⁺]_i was raised via the intracellular solution in the patch pipette, with an IC₅₀ of around 600 nM.

The mechanism by which Ca²⁺ inactivates TRPV4 remains unclear. Other TRPV channels (TRPV1 and TRPV6) have been shown to inactivate in a Ca²⁺-dependent fashion. TRPV6 shows biphasic inactivation kinetics, with a fast component in the millisecond and a slow component in the second range. The fast component was shown to depend on a sequence motif in the first intracellular loop [49], while the slow component is mediated by Ca²⁺-CaM binding to a C-terminal binding domain [46]. Both sites have no homologues in the TRPV4 sequence.

By analogy to TRPV1, however, for which a tyrosine residue in S6 has been shown to be involved in Ca²⁺ permeation and Ca²⁺-dependent desensitization [50], Watanabe et al. found that mutation of a phenylalanine residue at position 707 in S6 of TRPV4 (Figure 9.2) to alanine resulted in a reduction and slowing of current decay but did not modify the sensitivity to [Ca²⁺]_i. [48] In the same study, mutation of Glu797 in the C-terminal tail (Figure 9.2) increased spontaneous channel activity, but the increase in [Ca²⁺]_i in response to 4αPDD was similar [48]. Glu797 is close to the CaM-binding domain, and it remains to be seen whether the mutation of this residue influences CaM binding. Thus, Ca²⁺-dependent inactivation of TRPV4 seems to rely on structural features that have not yet been identified.

Interaction of Different Stimuli

Because a number of mechanisms converge on TRPV4, they are likely to interact and influence one another. Liedtke et al. found that responses of TRPV4 to decreases in osmolarity were larger at 37°C than at room temperature [12]. As described above for the individual activators, Gao et al. found that responses to most activators, including 4αPDD, PMA, hypotonic solutions, and shear stress, are increased at 37°C compared to room temperature [23]. For the response of TRPV4 to temperature, raising the osmolarity very significantly depressed responses to heat in HEK293 cells and oocytes, whereas decreasing the osmolarity initiated a response alone and potentiated the response to heat in HEK293 cells [19]. Somewhat surprisingly, however, the effects of 5,6-EETs, unlike the effects of hypotonic solutions [12,23], were found not to be affected by heat and were reduced in hypoosmotic conditions [51].

Biophysical Properties

TRPV1 to TRPV4 differ not only in their sequences but also in their biophysical properties from TRPV5 and TRPV6. Many of the differences in properties result from differences in the Ca²⁺ permeability of the isoforms. The IV relation of TRPV4 has a characteristic shape, showing outward rectification in Ca²⁺-containing solutions, but also an increase in conductance, and concomitantly in inward current, at negative membrane potentials (Figure 9.1B, C, D). The reversal potential in normal extracellular solutions is just positive to 0 mV (Figure 9.1C, D). The outward rectification results from a block by extracellular Ca²⁺. Lowering the extracellular [Ca²⁺] increases inward currents and leads to a loss of rectification in Ca²⁺-free solutions [52]. Asp672 and Asp682 in the pore region (Figure 9.2) are involved in the inhibition of TRPV4 currents by extracellular Ca²⁺. Neutralization of both leads to a strong reduction in inhibition and a linearization of the IV relation [52]. Interestingly, the IV shape is similar for single-channel and whole-cell currents. Because of the outward rectification, the single-channel conductance for outward currents (88–100 pS) is larger than that for inward currents (30–60 pS). A value of 310 pS at +80 mV, a value much larger than that in other studies, was obtained in a cursory investigation of single-channel properties in another study [12]. In studies other than the latter, the relatively large variability in the conductance values obtained for inward currents probably results from differences in the solutions bathing the external face of the patch. Lower values were obtained in more physiological Ca²⁺ concentrations [11] than in Ca²⁺-free solutions [37,41], as expected from the effect of Ca²⁺ on whole-cell currents [52]. Thus, where studied carefully, data on single-channel conductance are consistent, and variabilities in the conductance for inward currents result from the use of different experimental conditions.

The reversal potential of around +5 to +10 mV for TRPV4 with standard extra-and intracellular solutions and abolition of most of the inward current on replacement of extracellular cations with the large cation N-methyl-D-glucamine (NMDG⁺) indicates that the channel is a nonselective cation channel. TRPV4 is moderately Ca²⁺ permeable with P_Ca/P_Na values of between 6 and 10 [11,25]. Other divalent cations alsopermeate with a P_Mg/P_Na of 2–3, a P_Sr/P_Na of 9, and P_Ba/P_Na a of 0.7–7 [25,44,52]. The channel discriminates poorly among monovalent cations with P_K:P_Cs:P_Na:P_Li = 2:1.3:1:0.9 [25,35]. Mutations that affect Ca²⁺ inhibition and the shape of the IV relation also moderately reduce the Ca²⁺ permeability [52]. Mutation of Met680, which, by analogy to K⁺ channels, is located at the center of the selectivity filter (Figure 9.2), leads to a reduction in current and a strong reduction in Ca²⁺ permeability [52].

Blockers

Studies on trp channels suffer from the lack of specific pharmacological modulators. Like other members of the TRPV subfamily, TRPV4 is blocked by extracellular ruthenium red (RR) in micromolar concentrations [11,25,52]. The block by RR is voltage dependent, with inward currents being more strongly inhibited than outward currents. It is also notable that inward currents are inhibited much more rapidly than outward currents. For native currents mediated by TRPV4 in keratinocytes, but not in HEK293 cells, an increase in outward current on RR application has been reported [53]. In our hands, application of 10 μM RR to TRPV4 in HEK293 cells resulted in an increase in outward current followed by inhibition (unpublished observations). Neutralization of Asp682 (Figure 9.2), a residue involved in the block of TRPV4 by extracellular Ca²⁺ (see above), shifted the potential dependence of and slowed the block by RR [52]. Thus, Asp682 is likely to be in the outer region of the pore accessible to the large cation. TRPV4 is also inhibited by micromolar Gd³⁺¹¹ and SKF96365 [32], which, however, inhibit many other TRP channels and native cation channels in a similar concentration range and cannot be considered specific.