TRPM8 acute desensitization is mediated by calmodulin and requires PIP₂: distinction from tachyphylaxis

Abstract

The cold-sensing channel transient receptor potential melastatin 8 (TRPM8) features Ca²⁺-dependent downregulation, a cellular process underlying somatosensory accommodation in cold environments. The Ca²⁺-dependent functional downregulation of TRPM8 is manifested with two distinctive phases, acute desensitization and tachyphylaxis. Here we show in rat dorsal root ganglion neurons that TRPM8 acute desensitization critically depends on phosphatidylinositol 4,5-bisphosphate (PIP₂) availability rather than PIP₂ hydrolysis and is triggered by calmodulin activation. Tachyphylaxis, on the other hand, is mediated by phospholipase hydrolysis of PIP₂ and protein kinase C/phosphatase 1,2A. We further demonstrate that PIP₂ switches TRPM8 channel gating to a high-open probability state with short closed times. Ca²⁺-calmodulin reverses the effect of PIP₂, switching channel gating to a low-open probability state with long closed times. Thus, through gating modulation, Ca²⁺-calmodulin provides a mechanism to rapidly regulate TRPM8 functions in the somatosensory system.

transient receptor potential melastatin 8 (TRPM8) channel, the principal sensor of cold temperatures, is expressed in the somatosensory system and used to detect a broad range of cooling temperatures (Colburn et al. 2007; Dhaka et al. 2007; Kobayashi et al. 2005; McKemy 2005; McKemy et al. 2002; Nealen et al. 2003; Peier et al. 2002; Stucky et al. 2009). It may also be involved in pathological pain sensations under disease conditions (Colburn et al. 2007; Levine and Alessandri-Haber 2007; Xing et al. 2007). Interestingly, TRPM8 has been identified in respiratory, visceral, vascular, and cancer tissues, where it is implicated in respiratory and visceral disorders and cancer development (Cho et al. 2010; Hayashi et al. 2009; Knowlton and McKemy 2010; Mukerji et al. 2006; Sabnis et al. 2008; Tsavaler et al. 2001; Xing et al. 2008; Yang et al. 2006; Zhang and Barritt 2004). Thus TRPM8 function states and its regulation may have broad biological significances.

An important feature of TRPM8 channel is Ca²⁺-dependent functional downregulation following activation and subsequent increases of intracellular Ca²⁺ level (McKemy et al. 2002; Peier et al. 2002). This downregulation is manifested by a substantial reduction of TRPM8-mediated responses when cooling temperatures or TRPM8 agonists activate the channel (Daniels et al. 2009; Liu and Qin 2005; Premkumar et al. 2005; Rohacs et al. 2005). TRPM8 functional downregulation in the somatosensory system is believed to play a role in sensory adaptation to cold environments (Daniels et al. 2009; Liu and Qin 2005; Premkumar et al. 2005; Rohacs et al. 2005). One question about Ca²⁺-dependent downregulation of TRPM8 is what intracellular mechanisms are downstream of Ca²⁺. The answer to this question has significance in that physiological and pathological conditions may affect cold temperature sensitivity through the downstream mechanisms. Previous studies have shown that phosphatidylinositol 4,5-bisphosphate (PIP₂) hydrolysis following phospholipase C (PLC) activation results in TRPM8 downregulation (Daniels et al. 2009; Liu and Qin 2005; Rohacs et al. 2005). It has been suggested that Ca²⁺-dependent downregulation of TRPM8 may be partially due to the activation of Ca²⁺-sensitive PLC, which results in the hydrolysis of PIP₂ and subsequent TRPM8 functional downregulation (Daniels et al. 2009). In addition, protein kinase (PK)C and protein phosphatases have also been demonstrated to downregulate TRPM8 function (Abe et al. 2006; Premkumar et al. 2005). It has been suggested that Ca²⁺ entry following TRPM8 activation may activate Ca²⁺-sensitive PKC and protein phosphatases, which in turn results in Ca²⁺-dependent downregulation of TRPM8 (Abe et al. 2006; Premkumar et al. 2005). When TRPM8-mediated responses are carefully analyzed, two temporally distinct types of Ca²⁺-dependent downregulation can be identified. One is a rapid reduction of TRPM8-mediated response during a single application of agonist, and desensitization is a proper term for this type of functional downregulation. The other type of downregulation is manifested by the reduction of responses to repeated agonist applications, and tachyphylaxis is an appropriate term for this functional change of TRPM8. The presence of these two types of TRPM8 downregulation raises a new question as what specific roles PIP₂/PLC and PKC/protein phosphatase may play in each type of downregulation. In addition, most previous studies on TRPM8 regulation were performed in heterologous expression systems. It is necessary to directly examine TRPM8 regulation in cold-sensing sensory neurons because there are cellular and molecular differences between heterologous expression systems and sensory neurons.

Calmodulin (CaM) has been recognized to be an important intracellular signaling molecule to mediate Ca²⁺-dependent regulation of cation channels that are Ca²⁺ permeable on plasma membranes. These cation channels often become functionally downregulated when Ca²⁺-CaM binds to the channels. The functional downregulation of Ca²⁺-permeable cation channels by Ca²⁺-CaM serves as a negative feedback mechanism to prevent excessive Ca²⁺ influx into the cells to cause cell toxicity. CaM-mediated channel downregulation has been observed in a number of TRP channels, including heat-sensing channel TRPV1 (Lishko et al. 2007; Numazaki et al. 2003; Rosenbaum et al. 2004), warm temperature-sensing channel TRPV3 (Xiao et al. 2008), osmolarity-sensing channel TRPV4 (Strotmann et al. 2003), the highly Ca²⁺-selective channels TRPV5 and TRPV6 (Lambers et al. 2004; Niemeyer et al. 2001), and TRPM4 (Nilius et al. 2005). Although Ca²⁺-CaM may be a common mechanism to downregulate TRP channel function, the degree and kinetics of downregulation for each TRP channel may be substantially different. For example, TRPM4 and TRPM5, which are both activated directly by Ca²⁺, show a slow, Ca²⁺-dependent desensitization (Ullrich et al. 2005; Zhang et al. 2007). In addition to downregulation, Ca²⁺-dependent functional upregulation has also been observed in a number of TRP channels including TRPC5 (Ordaz et al. 2005), TRPV3 (Xiao et al. 2008), TRPV6 (Lambers et al. 2004), TRPA1 (Doerner et al. 2007; Zurborg et al. 2007), and TRPV4 (Strotmann et al. 2010). For TRPV4, a direct binding of Ca²⁺-CaM to the intracellular domain of the channel is shown to be critical for Ca²⁺-dependent functional upregulation (Strotmann et al. 2010). Thus Ca²⁺-CaM-mediated regulation of TRP channels may result in distinct functional changes for different TRP channels, and the differences could be further amplified by cell type differences. Although it is very likely that TRPM8 channels may be regulated by Ca²⁺-CaM, no study so far has provided direct experimental evidence to validate this idea. If TRPM8 channels are indeed regulated by Ca²⁺-CaM, there are questions that need to be further addressed. These questions include whether and how Ca²⁺-CaM may be involved in acute desensitization and/or tachyphylaxis and to what extent Ca²⁺-CaM may change the kinetics and magnitudes of TRPM8-mediated responses in sensory neurons. The answers to these questions will provide insights into cellular and molecular mechanisms of sensory processing of cold stimuli under physiological and pathological conditions. In the present study, we show in somatosensory neurons that multiple intracellular signaling pathways have distinct roles in acute desensitization and tachyphylaxis of TRPM8.

MATERIALS AND METHODS

Cell preparations.

For dorsal root ganglion (DRG) neurons, adult Sprague-Dawley rats (100–250 g, both sexes) were used. Animal care and use conformed to National Institutes of Health guidelines for care and use of experimental animals. Experimental protocols were approved by the University of Cincinnati Institutional Animal Care and Use Committee (ID: Gu 09-07-01-01). DRG neuron cultures were prepared as described previously (Tsuzuki et al. 2004). In brief, rats were deeply anesthetized with isoflurane (Henry Schein) and killed by decapitation. DRGs at cervical, thoracic, and lumbar levels were rapidly dissected out bilaterally in Leibovitz-15 medium (Mediatech) and incubated for 1 h at 37°C in minimum essential medium for suspension culture (S-MEM) (Invitrogen, Grand Island, NY) with 0.2% collagenase and 0.5% dispase and then triturated to dissociate neurons. The dissociated DRG neurons were then plated on glass coverslips precoated with poly-d-lysine (PDL, 12.5 μg/ml in distilled H₂O) and laminin [20 μg/ml in Hanks' balanced salt solution (HBSS); BD Biosciences] and maintained in MEM culture medium (Invitrogen) that also contained nerve growth factor (2.5 S NGF, 10 ng/ml; Roche Molecular Biochemicals, Indianapolis, IN), 5% heat-inactivated horse serum (JRH Biosciences, Lenexa, KS), uridine/5-fluoro-2′-deoxyuridine (10 μM), 8 mg/ml glucose, and 1% vitamin solution (Invitrogen). The cultures were maintained in an incubator at 37°C with a humidified atmosphere of 95% air and 5% CO₂. Unless otherwise indicated, cells were used within 72 h after plating.

The human embryonic kidney (HEK) 293 cell line was obtained from American Type Culture Collection (Manassas, VA). The cells were cultured in DMEM (Hyclone) supplemented with 10% FBS (Hyclone) and 0.5% penicillin-streptomycin in culture flasks. Wild-type mouse TRPM8 was stably transfected into the HEK 293 cells with the protocol described in a previous study (Hu et al. 2004). For electrophysiology recordings, cells were plated on PDL-coated glass coverslips 1 day before being used for experiments.

Electrophysiology.

Patch-clamp recordings were performed on menthol-sensitive DRG neurons and HEK 293 cells stably expressing TRPM8. Menthol-sensitive neurons were preidentified with Ca²⁺ imaging (Xing et al. 2007), and cells that showed strong Ca²⁺ responses after menthol application (100 μM, 5 s) were used for recordings. The duration of time elapsed between the application of menthol to identify menthol-sensitive neurons with calcium imaging and electrophysiological recordings was 20–30 min. Cells were perfused at room temperature of 24°C with normal bath solution containing (in mM) 150 NaCl, 5 KCl, 2 MgCl₂, 2 CaCl₂, 10 glucose, and 10 HEPES, pH 7.3 adjusted with NaOH and osmolarity 320 mosM adjusted with sucrose. In some experiments, calcium was omitted from the bath solution and 5 mM EGTA was added for extracellular Ca²⁺-free experiments. Unless otherwise indicated, cells were voltage-clamped at −70 mV in the whole cell configuration with a Multiclamp 700A amplifier (Axon Instruments), and signals were sampled at 5 kHz and filtered at 2 KHz with pCLAMP 9.0 (Axon Instruments). At this negative holding potential and the temperature of 24°C, there was little basal activation of TRPM8. Electrode internal solutions contained (in mM) 110 Cs₂SO₄, 2.4 MgCl₂, 10.0 HEPES, 5.0 Na₂ATP, and 0.33 GTP-Tris salt; pH was adjusted to 7.35 with NaOH and osmolarity was adjusted to 320 mosM with sucrose. In some experiments to test intracellular Ca²⁺ buffer effects, electrode internal solutions also contained 5 mM EGTA or 5 mM BAPTA. Recording electrode resistance was ∼5.0 MΩ. Whole cell membrane currents were evoked by rapidly applying menthol solution (100 μM) or cold bath solution (12 ± 0.5°C) to cells through a glass tube (500-μm ID) positioned 1.0 mm away from cells. The on-time for menthol solution application to a cell under recording was usually <500 ms. To achieve rapid cooling on recorded cells, bath solution was precooled and maintained at a constant low temperature inside the temperature controlling head stage of the TCM-1 temperature controlling system (Warner instruments). The cold bath solution was rapidly applied (2 ml/min) to the cells from a short tube (0.2-cm L, 500-μm ID). With this setting, the on-time to reach stable cooling of 12 ± 0.5°C usually took <3 s, much faster than the time of temperature drop (>30 s) in most previous studies (Daniels et al. 2009; McKemy et al. 2002; Peier et al. 2002). Unless otherwise indicated, other testing compounds were directly introduced into the cell by mixing them in the electrode internal solutions.

For cell-attached and inside-out experiments, external bath and electrode internal solutions were ionically symmetrical and contained (in mM) 140 Na-gluconate, 10 NaCl, and 10.0 HEPES; pH was adjusted to 7.35 with NaOH and osmolarity adjusted to 320 mosM with sucrose. Recording electrode resistance was ∼10 MΩ. In inside-out experiments in which cold response was examined, perfusing solution was maintained at 15°C. For experiments in which testing agents were applied, excised membranes were quickly transferred to a separate chamber at 24°C. Experiments were performed with membrane potential held from −100 mV to 100 mV with an Axopatch 200B amplifier. Signals were sampled at 10 kHz and filtered at 2 kHz. Data were collected with Clampex 9 software.

Phosphatidylinositol-4,5-bisphosphate C-8 (sodium salt) (PIP₂) was purchased from Cayman Chemical; it was dissolved in electrode internal solution, flash frozen, stored at −20°C, and used within 48 h. CaM (porcine) was purchased from A.G. Scientific. Ophiobolin A was purchased from Santa Cruz Biotechnology. Okadaic acid, U73122, Go6976, and KN-62 were purchased from Tocris. Menthol, allyl isothiocyanate (AITC), bisindolylmaleimide (BIM), BAPTA, and EGTA were purchased from Sigma. Final concentrations of compounds used were PIP₂ 50 μM (Rohacs et al. 2005), CaM 10 μM (Black et al. 2004), ophiobolin A 100 μM (Leung et al. 1985), okadaic acid 100 nM (Schonthal 1998), KN-62 25 μM, menthol 100 μM, AITC 100 μM, and BIM 1 μM (Sarria and Gu 2010), and Go6976 1 μM (Gschwendt et al. 1996).

Unless otherwise indicated, these compounds were included in the electrode internal solution and applied into cells through recording electrodes. Recordings were performed after 4-min equilibrium of intracellular contents with electrode internal solutions.

Data analysis.

Clampfit 9 was used for analysis of whole cell and unitary currents and least-square fittings and to create figures. For acute desensitization, percentage of desensitization was obtained by dividing the current reduction at 30 s by the peak current value. For tachyphylaxis, percentage of tachyphylaxis was obtained by dividing current reduction between peaks of two stimuli by peak current values of the first stimulus. For cell-attached and inside-out recordings, channel openings and closings were determined off-line by using a 50% crossing threshold. The probability of channel opening (P_o) was calculated as the ratio of the total channel open time to total time. For single-channel recordings, fitting of the closed and open dwell time components were obtained by using the variable metric-maximum likelihood method. The usefulness of adding exponential components was assessed by using an F-test (De Koninck and Mody 1994) and also by visual inspection. Unless otherwise specified, data in all figures are presented as means ± SE. Analysis of variance (ANOVA, 1-way) was used for statistical analyses of data sets of multiple groups and/or treatments followed by Student-Newman-Keuls post hoc test for all pairwise comparisons or Dunnett post hoc test for comparison against control group. Student's t-test was used to evaluate the significance of changes in mean values between two groups. Statistical significance was considered at the P < 0.05 level.

RESULTS

Acute desensitization and tachyphylaxis of TRPM8 in rat DRG neurons.

We performed whole cell recordings on rat DRG neurons that were voltage-clamped at −70 mV and bathed in an external solution containing 2 mM Ca²⁺ at the room temperature of 24°C. When menthol (100 μM, 30 s) was applied for the first time, inward currents showed acute desensitization (Fig. 1A, left) with current amplitude reduced by 43.8 ± 5.9% (n = 14) at the end of 30-s menthol application (Fig. 1C). For tachyphylaxis, menthol was applied a second time after a 10-min interval and maximal inward currents were reduced by 69.0 ± 3.9% (n = 12, Fig. 1, A, right, and D) compared with the peak currents of the first menthol application. Similar to menthol, both acute desensitization and tachyphylaxis were also observed when a cold bath solution was rapidly applied to menthol-sensitive DRG cells preidentified with Ca²⁺ imaging. A rapid application of cold bath solution (12 ± 0.5°C) for 30 s resulted in acute desensitization by 58.6.3 ± 9.3% (n = 9, Fig. 1, B, left, and C) at the end of cold bath application. A rapid temperature drop rather than a slow temperature ramp was found to be essential to observe cold-induced acute desensitization. In this experiment, the temperature drop from 24°C to 12°C was <3 s at recording sites, more than 10 times faster than previous studies using temperature ramp (McKemy et al. 2002; Thut et al. 2003). Repeated cold bath applications with a 10-min interval resulted in tachyphylaxis by 62.2 ± 11.8% (n = 7, Fig. 1, B, right, and D) for the second stimulation.

Fig. 1.

Menthol- and cold-evoked currents in dorsal root ganglion (DRG) neurons bath-perfused at room temperature. A: sample traces show whole cell currents recorded from a rat DRG neuron after 2 applications of 100 μM menthol. B: sample traces show whole cell currents recorded from a rat DRG neuron after 2 applications of a cold bath solution. Two traces under the currents show the temperature step that produces rapid cooling from room temperature of 24°C to 12°C. In both A and B, each application was 30 s and intervals between each application were 10 min. p, Peak current of 1st application; d, end of 1st application; t, current maximum of 2nd application. C: pooled results of acute desensitization during the 1st application of 100 μM menthol (n = 9) or 12°C cold bath solution (n = 7). D: pooled results of tachyphylaxis following the consecutive application of 100 μM menthol (n = 9) or 12°C cold bath solution (n = 7). In all experiments cells were held at −70 mV. Data represent means ± SE.

We determined whether both acute desensitization and tachyphylaxis occurred in a similar manner when DRG cells were bath perfused at a more physiological temperature (32°C) before the application of menthol and cold stimulation. Under this condition, acute desensitization occurred with peak current amplitude being reduced by 57.79 ± 7.43% (n = 10) at the end of 30-s menthol application (Fig. 2, A, left, and C). Tachyphylaxis occurred with maximal inward currents of the second menthol application being reduced by 68.28 ± 7.25% (n = 8; Fig. 2, A, right, and D) compared with the peak currents of the first menthol application. Similar to menthol, both acute desensitization and tachyphylaxis were also observed when a cold bath solution was rapidly applied to menthol-sensitive DRG cells preidentified with Ca²⁺ imaging. A rapid application of cold bath solution (12 ± 0.5°C) for 30 s resulted in acute desensitization by 48.21 ± 6.39% (n = 12; Fig. 2, B, left, and C) at the end of cold bath application. Repeated cold bath applications with a 10-min interval resulted in tachyphylaxis by 71.86 ± 4.2% (n = 8; Fig. 2, B, right, and D) for the second stimulation. Thus both acute desensitization and tachyphylaxis occurred in a similar manner under the conditions when cells were bath-perfused either at room temperature or at more physiological temperature.

Fig. 2.

Menthol- and cold-induced responses in DRG neurons bath-perfused at 32°C. A: sample traces show whole-cell currents recorded from a rat DRG neuron after 2 applications of 100 μM menthol. The menthol solution was at the room temperature of 24°C. B: sample traces show whole cell currents recorded from a rat DRG neuron after 2 applications of a cold bath solution. Two traces under the currents show the temperature step that produces rapid cooling from 32°C to 12°C. In both A and B, each application was 30 s and intervals between each application were 10 min. C: pooled results of acute desensitization during the 1st application of 100 μM menthol (n = 8) or 12°C cold bath solution (n = 10). D: pooled results of tachyphylaxis following the consecutive application of 100 μM menthol (n = 8) or cold 12°C cold bath solution (n = 10). In all experiments cells were held at −70 mV. Data represent means ± SE.

Similar to DRG neurons, HEK 293 cells expressing mouse TRPM8 displayed an acute desensitization to menthol application (100 μM, 30 s) in which peak currents were reduced by 47.56 ± 4.61% (n = 8). Tachyphylaxis occurred with maximal inward currents of the second menthol application being reduced by 71.25 ± 5.34% (n = 8). When cold (12 ± 0.5°C, 30s) was used as an agonist, acute desensitization and tachyphylaxis were 63.30 ± 3.33% and 65.77 ± 8.57%, respectively (n = 13). We have shown previously that TRPM8 rather than TRPA1 accounts for strong responses evoked by 100 μM menthol in rat DRG neurons (Sarria and Gu 2010). Nevertheless, menthol was shown to have a promiscuous relationship with TRPA1 receptors in expression systems (Macpherson et al. 2006). Consistent with our previous finding, the TRPA1 agonist AITC did not evoke detectable inward currents in menthol- and cold-sensitive cells in this study (100 μM, n = 5/5).

PIP₂ attenuates tachyphylaxis but not acute desensitization of TRPM8.

Do acute desensitization and tachyphylaxis have the same or distinctive molecular mechanisms? To answer this question, we set out to examine how PIP₂ availability affected acute desensitization and tachyphylaxis in DRG neurons, since PIP₂ has been indicated to play an important role in regulating TRPM8 function (Liu and Qin 2005; Rohacs et al. 2005). When the PLC inhibitor U73122 (25 μM) was included in the recording internal solution to block PIP₂ hydrolysis, acute desensitization of menthol-evoked currents was 36.6 ± 8.0% (n = 9; Fig. 3, B and D) and was as strong as the control, while tachyphylaxis was significantly attenuated to 40.2 ± 9.4% (n = 9; Fig. 3, B and E). A possible explanation for the lack of effect by U73122 on acute desensitization could be incomplete block of PLC. To test this possibility we next examined whether direct application of PIP₂ (50 μM) into cells would prevent acute desensitization along with tachyphylaxis. To our surprise, direct application of PIP₂ into cells did not prevent acute desensitization at all. On the contrary, the acute desensitization became even more substantial at the degree of 70.0 ± 6.7% (n = 9) and also reproducible upon repeated menthol applications (Fig. 3, C and D). Tachyphylaxis, on the other hand, was greatly attenuated by PIP₂ to only 32.7 ± 10.7% (n = 9; Fig. 3, C and E). Neither U73122 nor PIP₂ had significant effect on the peak current amplitude evoked by the first menthol application (Fig. 3F). The opposite effects of PIP₂ on acute desensitization and tachyphylaxis suggest that the two regulatory processes may be mechanistically different.

Fig. 3.

Phosphatidylinositol 4,5-bisphosphate (PIP₂) attenuates tachyphylaxis and confers transient receptor potential melastatin 8 (TRPM8) functional state for acute desensitization. Sample traces show menthol-evoked whole cell currents from rat DRG neurons of untreated (A, control), phospholipase C (PLC) blocker U73122-treated (B), and PIP₂-treated (C) cells. In each group, menthol (100 μM) was applied twice at an interval of 10 min and the duration of each application was 30 s. D: summary of desensitization from peak (p) to the end (d) of 1st menthol application in untreated (n = 12) and U73122 (n = 9)- and PIP₂ (n = 9)-treated cells. E: summary of tachyphylaxis from the peak current (p) of the 1st menthol application to the current maximum of the 2nd menthol application (t) for untreated (n = 12) and U73122 (n = 9)- and PIP₂ (n = 9)-treated cells. F: average peak current of the 1st menthol application for the 3 groups; n is the same as in D and E. U73122 (25 μM) and PIP₂ (50 μM) were included in the internal electrode solution, which in all cases was allowed to dialyze for 3–4 min before application of menthol (100 μM). All experiments were carried out at 24°C with cells held at −70 mV, and the bath solution contained 2 mM Ca²⁺. Data represent means ± SE. *P < 0.05 vs. control, 1-way ANOVA and Dunnett post hoc.

Both acute desensitization and tachyphylaxis are Ca²⁺ dependent, but they have different Ca²⁺ sensitivity.

Because both acute desensitization and tachyphylaxis are activity dependent, we next tested whether both processes are Ca²⁺ dependent and, if so, whether the two processes have different Ca²⁺ sensitivity. We compared menthol-evoked currents in DRG cells under the following conditions: 1) without including a Ca²⁺ buffer in the internal solution, 2) including the slow Ca²⁺ buffer EGTA (5 mM) or 3) the fast Ca²⁺ buffer BAPTA (5 mM) in the internal solution, or 4) including BAPTA (5 mM) in the internal solution together with an extracellular Ca²⁺-free condition. In the presence of 5 mM EGTA, acute desensitization (Fig. 4, A and D) was 42.4 ± 5.6% (n = 11) and tachyphylaxis (Fig. 4, A and E) was 56.9 ± 6.9% (n = 8), not significantly different from those without a Ca²⁺ buffer (Fig. 3, D and E). Interestingly, 5 mM BAPTA in the internal solution significantly attenuated tachyphylaxis (28.0 ± 6.8%, n = 9; Fig. 4, B and E) but had no significant effect on acute desensitization (31.1 ± 7.6%, n = 11; Fig. 4, B and D). Further removal of extracellular Ca²⁺, however, significantly attenuated both acute desensitization (7.7 ± 3.6%, n = 10; Fig. 4, C and D) and tachyphylaxis (18.4 ± 4.8%, n = 7; Fig. 4, C and E). There was no significant difference in peak currents of the first menthol-evoked responses under different Ca²⁺ buffering conditions (Fig. 4F). Thus, while both acute desensitization and tachyphylaxis are Ca²⁺ dependent, acute desensitization is more sensitive to Ca²⁺ compared with tachyphylaxis, as is evidenced by the differential effect of the fast Ca²⁺ buffer BAPTA.

Fig. 4.

Both acute desensitization and tachyphylaxis are Ca²⁺ dependent but show different Ca²⁺ sensitivity. A–C: sample traces show menthol-evoked currents from rat DRG neurons under the following intracellular Ca²⁺ buffering conditions: 5 mM EGTA (A), 5 mM BAPTA (B), 5 mM BAPTA in a Ca²⁺-free (0 Ca²⁺) bath solution (C). In each group, menthol (100 μM) was applied twice at an interval of 10 min and the duration of each application was 30 s. D: desensitization from the peak current (p) to the end (d) of the 1st menthol application for EGTA (n = 11), BAPTA (n = 10), and BAPTA/0 Ca²⁺ (n = 10) groups. E: average tachyphylaxis from the peak current of the 1st menthol application (p) to the maximum current of the 2nd menthol application (t) in EGTA (n = 8), BAPTA (n = 9), and BAPTA/0 Ca²⁺ (n = 7) groups. F: average peak current amplitude of the 1st menthol application for the 3 groups; cell numbers are the same as in D. All experiments were carried out at 24°C with cells held at −70 mV. Calcium chelators were introduced into the cells via the recording internal solution and allowed to dialyze for 3–4 min before menthol application. Data represent means ± SE. *P < 0.05 vs. control, 1-way ANOVA and Dunnett post hoc.

Inhibiting protein kinase C or protein phosphatase 1,2A attenuates tachyphylaxis but not acute desensitization.

Previous studies showed that activation of either PKCs or protein phosphatases downregulated TRPM8 function (Abe et al. 2006; Premkumar et al. 2005). We set out to determine whether acute desensitization in DRG cells may be mediated by PKC/protein phosphatases. To test this possibility, we examined Ca²⁺-dependent acute desensitization in the presence of the PKC inhibitor BIM or in the presence of the protein phosphatase 1,2A (PP1,2A) inhibitor okadaic acid. These experiments were conducted under the condition that intracellular Ca²⁺ was not buffered so that tachyphylaxis was maximized. Under this condition and in the presence of 1 μM BIM (Fig. 5, B and F) or 100 nM okadaic acid (Fig. 5, C and F), tachyphylaxis was 39.7 ± 8.6% (n = 9) and 40.7 ± 8.1% (n = 9), respectively, significantly less than the control without the inhibitors (Fig. 5, A and F). On the other hand, acute desensitization remained strong and was not significantly affected by either inhibitor in the first menthol application (Fig. 5, B, C, and F). Acute desensitization reoccurred in subsequent menthol application when PKC or PP1,2A was blocked (Fig. 5, B and C, compared with Fig. 5A). Acute desensitization in the second menthol application was 52.0 ± 7.8% (n = 8) in the presence of BIM and 44.5 ± 11.5% (n = 7) in the presence of okadaic acid, not significantly different from the degree of acute desensitization during the first menthol application (43.8 ± 5.9%, n = 14). Thus inhibition of PKC and protein phosphatase 1,2A did not prevent acute desensitization, although it attenuated tachyphylaxis.

Fig. 5.

Inhibition of calmodulin attenuates acute desensitization; blocking protein kinase C (PKC) and protein phosphatase 1,2A reduces tachyphylaxis. A–E: sample traces show menthol-evoked currents from rat DRG neurons under the following conditions: control (A), 1 μM PKC blocker bisindolylmaleimide (BIM, B), 100 nM protein phosphatase 1,2A blocker okadaic acid (OA, C), 100 μM calmodulin blocker ophiobolin A (Oph, D), 100 μM ophiobolin A + 50 μM PIP₂ (E). F: summary of acute desensitization (left) and tachyphylaxis (right) in control (n = 14) and with BIM (n = 9), OA (n = 9), Oph (n = 8), and PIP₂ + Oph (n = 10). In each group, menthol (100 μM) was applied twice at an interval of 10 min and the duration of each application was 30 s. Data represent means ± SE. In all experiments, currents were recorded at a holding potential of −70 mV in bath solution containing 2 mM Ca²⁺ and a temperature of 24°C. Inhibitors and PIP₂ were introduced into the cells via the recording internal solution and allowed to dialyze for 3–4 min before menthol application (100 μM). *P < 0.05 vs. control, 1-way ANOVA and Dunnett post hoc.

Blocking calmodulin diminishes acute desensitization.

CaM has been shown to directly regulate functions of some TRP channels (Nilius et al. 2005; Numazaki et al. 2003; Zhu 2005). We asked whether CaM is involved in regulating TRPM8 in DRG cells and, if so, whether acute desensitization and/or tachyphylaxis is/are mediated by calmodulin. To test this idea, we determined whether ophiobolin A, a specific CaM inhibitor, affected acute desensitization of TRPM8. In the presence of 100 μM ophiobolin A, acute desensitization was substantially attenuated to 21.2 ± 4.7% (n = 8; Fig. 5, D and F), while tachyphylaxis remained strong (62.33 ± 12.31%, n = 8) and was not significantly attenuated (Fig. 5, D and F). Since PIP₂ is essential for the appearance of desensitization current upon repeated stimulation (Fig. 3C), we included both ophiobolin A (100 μM) and PIP₂ (50 μM) in the recording solutions. Under this condition both acute desensitization and tachyphylaxis were greatly diminished to 19.6 ± 6.5% (n = 10) and 11.3 ± 11.2% (n = 8), respectively (Fig. 5, E and F). Acute desensitization was unlikely mediated by CaMKII or PKD since CaMKII inhibitor KN-62 (25 μM) or PKD inhibitor Go6976 (1 μM) did not significantly attenuate acute desensitization [KN-62 39.01 ± 6.42% (n = 6), Go6976 49.45 ± 7.34% (n = 6) vs. control 43.8 ± 5.9% (n = 14)]. These results support our hypothesis that CaM is a primary cause of acute desensitization and PIP₂ hydrolysis, on the other hand, is an important cause of tachyphylaxis.

PIP₂ and calmodulin inversely modulate TRPM8 channel gating.

To provide further insights into mechanisms by which PIP₂ and CaM regulate TRPM8 channels, we performed single-channel recordings from membranes of HEK 293 cells that stably express TRPM8 channels and studied effects of PIP₂ and CaM on TRPM8 gating properties. In inside-out patches, unitary currents were observed in TRPM8-expressing HEK 293 cells (Fig. 6A) when 100 μM menthol was included in patch electrodes or temperature dropped to 15°C. Unitary currents of membrane patches showed a lack of outward rectification, and single-channel conductances were 59.0 ± 2.9 pS with menthol (n = 7, 10, and 5 for each holding potential at 100, 70, and 40 mV, respectively) and 62.0 ± 3.1 pS (n = 5) measured at 15°C (Fig. 6B). TRPM8 channel shows voltage-dependent increases in P_o with either 100 μM menthol (n = 7, 10, 5, 7, and 5 for each holding potential at 100, 70, 40, −70, and −100 mV, respectively) or 15°C (n = 5) (Fig. 6C). The voltage-dependent change of channel P_o mirrors the outward rectification property observed in whole cell TRPM8 currents described previously (McKemy et al. 2002). With either menthol (n = 7, 10, 5, 7, and 5 for each holding potential at 100, 70, 40, −70, and −100 mV, respectively) or cold stimulus (n = 5), there was a slight increase in the mean open time as voltage increased (Fig. 6D). The unitary currents were not observed in control HEK 293 cells that did not express TRPM8. In cell-attached configuration before excision of the patched membranes, the channel P_o was high at 0.24 ± 0.04 and 0.15 ± 0.05 when cells were held at 100 mV or 70 mV, respectively (n = 6; Fig. 7, A–C). After membranes were excised and inside-out configuration was acquired, channel P_o at 70 mV was significantly reduced to 0.02 ± 0.01 (n = 6, Fig. 7, A–C). When PIP₂ (50 μM) was added in the solution, P_o increased to 0.20 ± 0.05 (n = 6), restored to the cell-attached values (Fig. 7, A–C). Exposure of the excised membrane to PIP₂ did not significantly change the amplitude of unitary currents (n = 7, Fig. 7D). Channel conductance calculated from unitary currents was 53.0 ± 4.7 pS (n = 6) without PIP₂ and 56.0 ± 1.8 pS (n = 7) with PIP₂. Mean open time was also not affected (Fig. 7E), but mean closed time was significantly shortened from 107.9 ± 19.4 ms (n = 6) in the absence of PIP₂ to 7.2 ± 1.6 ms when PIP₂ was present (Fig. 7F). While PIP₂ increased channel P_o to 0.24 ± 0.07 in yet another group of cells, addition of CaM (10 μM) along with 40 μM Ca²⁺ reversed the effects caused by PIP₂, resulting in a significant reduction of P_o to 0.05 ± 0.02 (n = 7; Fig. 8, A–C). Addition of calcium (40 μM) alone did not have significant effects. Ca²⁺-CaM did not have a significant effect on the amplitude of unitary currents measured at 70 mV (Fig. 8D), and the calculated conductance of 54 ± 2.6 pS (n = 7) was not significantly different from that in the absence of Ca²⁺-CaM. When PIP₂ or Ca²⁺-CaM was added, mean open time was not significantly different from the control group (Fig. 8E). However, mean closed time was significantly shortened from 96.6 ± 14.5 ms (n = 7) in the control to 10.16 ± 3.15 ms (n = 7) with PIP₂. Addition of Ca²⁺-CaM reversed the mean closed time to 86.7 ± 25.6 ms (Fig. 8F; n = 7). Dwell time of channel open could be best fitted into a two-exponential model, and closed time was best fitted into three components for control, with PIP₂, and PIP₂ plus CaM (Fig. 8F). There were no changes of relative portions of open time for either term (Fig. 8, F and G, left). On the other hand, the portions of closed time showed a significant shift from longer closing (3rd term) to shorter closing (1st and 2nd terms) when PIP₂ was added (Fig. 8, F, right, and G, right). Addition of Ca²⁺-CaM reversed the channel closed-time distribution and shifted it back to the pre-PIP₂ values with longer closings occupying 45 ± 11% of the time versus 17.1 ± 1.8% when only PIP₂ was present (n = 7, Fig. 8, F and G, right). Thus PIP₂ and Ca²⁺-CaM inversely regulate TRPM8 gating.

Fig. 6.

Single-channel characteristics of TRPM8. A: unitary currents recorded from an inside-out patch with either electrode containing 100 μM menthol (left) or in a cold bath solution (15°C, no menthol, right). Holding potentials were −70, 70, and 100 mV. Closed (c) and open (o) states are indicated. B: current (I)-voltage (V) relationship of TRPM8 unitary currents and calculated conductance (G). C: open probability (P_o) at different holding potentials. D: dwell time of channel open at different holding potentials. In B–D, solid circles represent tests of menthol and open circles tests of cold; n = 7, 10, 5, 7, and 5 for each holding potential at 100, 70, 40, −70, and −100 mV, respectively. HEK 293 cells stably expressing TRPM8 were used.

Fig. 7.

Modulation of TRPM8 channel gating by PIP₂. A: unitary currents recorded from a single patch under the following conditions, from top to bottom: cell-attached configuration (C/A) with voltage clamped at 70 mV and 100 mV and inside-out configuration (I/O) with voltage clamped at 70 mV before and after addition of 50 μM PIP₂. Open (o) and closed (c) states are indicated in the 1st trace. B: time course of TRPM8 channel P_o for the recording illustrated in A. Time bins: 2 s. Breaks: 15 s. C: mean P_o. D: mean unitary current amplitude. E: mean open time. F: mean closed time. In all experiments, HEK 293 cells stably expressing mouse TRPM8 were used and 100 μM menthol was included in the recording electrodes. Data represent means ± SE; n = 6 for each group. *P < 0.05 vs. control, 1-way ANOVA and post hoc Student-Newman-Keuls (STK) pairwise comparisons.

Fig. 8.

Differential modulation of TRPM8 gating by PIP₂ and calmodulin (CaM). A: unitary currents recorded from an inside-out patch held at 70 mV before (control; top), during application of 50 μM PIP₂ (middle), and after addition of 10 μM CaM + 40 μM Ca²⁺ (bottom). B: time course of TRPM8 P_o for the recording illustrated in A. Time bin: 2 s. Breaks: 15 s. C: mean P_o at 70 mV in control (n = 7), after PIP₂ (n = 7), and after Ca²⁺-CaM (n = 7). D: mean unitary current amplitude. E: mean open and closed time (τ). F: dwell time histograms of open and closed states in control (top), after PIP₂ (middle), and after addition of Ca²⁺-CaM (bottom). Curves show the variable metric-maximum likelihood fit; n is the number of events. There were 2 open components and 3 closed components. τ, Time constant. Percentage values indicate relative areas of each component. In average, τ1 = 0.37 ± 0.05 ms and τ2 = 1.7 ± 0.26 ms for open time. For closed time, τ1 = 0.68 ± 0.05 ms, τ2 = 6.3 ± 0.92 ms, and τ3 = 117 ± 31 ms. G: summary of the relative areas occupied by each component in the open state (left) and the closed state (right) during control, PIP₂, and PIP₂ + Ca²⁺-CaM. Data represent means ± SE. HEK 293 cells stably expressing TRPM8 were used, and 100 μM menthol was included in recording electrodes. *P < 0.05 vs. control, 1-way ANOVA and post hoc STK pairwise comparisons.

DISCUSSION

In the present study we have investigated the two phases of Ca²⁺-dependent downregulation of TRPM8 channels. We have shown for the first time that acute desensitization is triggered by Ca²⁺-CaM and that PIP₂ availability is essential for the appearance of acute desensitization. We have provided clear evidence that PIP₂ hydrolysis and PKC/PP1,2A activity are two main causes of tachyphylaxis. We have also demonstrated that Ca²⁺-CaM and PIP₂ inversely modulate TRPM8 gating to confer functional states of TRPM8 channel for acute desensitization and tachyphylaxis.

The degree of acute desensitization and tachyphylaxis to consecutive TRPM8 activations has had large variations among different studies (Malkia et al. 2009; McKemy et al. 2002; Rohacs et al. 2005; Thut et al. 2003). Several factors may account for the large variations, including on-time speed of cooling, TRPM8 agonists applied to cells, techniques used for measuring TRPM8 activation, and cell types. High on-time speed of an agonist is critical for revealing rapid desensitization of any ion channel. We have noted that the phase of acute desensitization would be small or even unnoticeable when menthol or cold solutions were not applied fast enough (on-time speed is slow). The on-time speed for cold application in most of the previous studies was probably too slow (>30 s) to reveal acute desensitization and its kinetics (McKemy et al. 2002; Thut et al. 2003). Previous studies using Ca²⁺ imaging technique also yielded the appearance of slow TRPM8 desensitization/tachyphylaxis (Malkia et al. 2009; Thut et al. 2003) because of its technical limitations. Different cell types may also contribute to the variations in the degree of acute desensitization and tachyphylaxis. The kinetics of acute desensitization for TRPM8 appeared to be slower in very big cells such as oocytes (McKemy et al. 2002), probably because agonist on-time could not be achieved fast and uniformly for very big cells. Substantial recovery of TRPM8 from tachyphylaxis was reported in oocytes when the cells were held at 32°C (Daniels et al. 2009). It was thought that physiological temperature might favor PIP₂ resynthesis, thereby allowing a faster recovery of TRPM8 from functional downregulation. In our study on DRG neurons and under a whole cell recording configuration, tachyphylaxis of TRPM8 currents was profound with little recovery in 10 min when cells were held at either room temperature or 32°C. The lack of recovery from tachyphylaxis in our DRG neurons could be due to a severe dialysis of some intracellular components that are essential for PIP₂ resynthesis. Consistent with this idea, tachyphylaxis of TRPM8 was largely prevented when PIP₂ was included in our recording solution. Dialysis of other intracellular components (Andersson et al. 2007; Zakharian et al. 2009) may also account for some degree of TRPM8 tachyphylaxis and poor recovery. It is conceivable that TRPM8 tachyphylaxis in sensory neurons may be less severe, and its recovery may be faster in intact cells under physiological conditions.

An intriguing finding in our study is that inhibiting PLC to reduce PIP₂ metabolism or directly introducing PIP₂ into cells results in a higher degree of acute desensitization that reoccurs upon repeated TRPM8 activation. Previous studies have indicated that PIP₂ is essential for maintaining functions of TRP channels including TRPM8 (Karashima et al. 2008; Liu and Qin 2005; Liu and Liman 2003; Lukacs et al. 2007; Rohacs et al. 2005; Zhang et al. 2005). Our study indicates that PIP₂ maintains TRPM8 in sensory neurons at a high functional state that is manifested by a high P_o due to the promotion of short channel closings. This is consistent with a recent study of TRPM8 gating by PIP₂ using planar lipid bilayers (Zakharian et al. 2010). The high functional state is a prerequisite for TRPM8 to undergo gating switch to the low functional state, which is characterized by low P_o dominated by long closed times in single-channel activities. In acute desensitization, Ca²⁺-CaM plays a key role in mediating such a gating switch (Fig. 7).

Acute desensitization of TRPM8 is highly sensitive to Ca²⁺ near the channel pore, as evidenced by its resistance to fast intracellular Ca²⁺ buffering by BAPTA. In contrast, tachyphylaxis may require Ca²⁺ diffusion, as it can be effectively attenuated by intracellular BAPTA. The different sensitivity to Ca²⁺ buffering may be due to differences in calcium affinities between CaM and PLC (Chin and Means 2000; Wiesner et al. 1996), which would favor a sharper acute desensitization by CaM and a kinetically slower tachyphylaxis through PIP₂ hydrolysis by PLC. The proximity of their regulatory sites to the channel pore may be another factor underlying different Ca²⁺ buffering sensitivity. Our results suggest that the site involved in acute desensitization either has a very high Ca²⁺ affinity or is located on or very close to intracellular termini near the pore of the TRPM8 channels. In our quest to find out whether acute TRPM8 desensitization is directly mediated by Ca²⁺-CaM or through other linked intracellular signaling pathways, we observed that neither PKC or phosphatase inhibitors attenuated acute desensitization. Additionally, the degree of acute desensitization was not affected by either the CaMKII inhibitor KN-62 [39.01 ± 6.42% (n = 6)] or the PKD inhibitor Go6976 [49.45 ± 7.34% (n = 6)] compared with control [43.8 ± 5.9% (n = 14)]. These results, together with direct effects of CaM on TRPM8 gating in inside-out recordings, point to a direct action of CaM during acute desensitization. This is consistent with kinetics and high Ca²⁺ sensitivity of acute desensitization. Binding sites for CaM have been identified in several other TRP channels including TRPM4 and TRPV1, and Ca²⁺-CaM binding at these sites is suggested to confer Ca²⁺-dependent regulation of these channels (Nilius et al. 2005; Numazaki et al. 2003). A CaM binding site has been located at amino acids 145–198 of the NH₂ terminus of TRPM8 with a pull-down assay (Qin and Flores 2007). This raises the possibility that the binding of CaM at this site may cause the channel to enter the low functional state, resulting in acute desensitization (Fig. 9). Future studies using point mutation are needed to further pinpoint the sites that are critical for CaM binding and acute desensitization.

Fig. 9.

Schematic diagram of Ca²⁺-dependent modulation of TRPM8 channels. Ca²⁺ entry through TRPM8 channels or other sources can activate CaM, which leads to a switch of TRPM8 channels from a high to a low activity state, thereby causing acute desensitization at the macroscopic current level. CaM binding sites may be located at NH₂ terminus of TRPM8 channels. CaM-mediated acute desensitization may not be observed if TRPM8 channels are already at low activity state because of other regulatory mechanisms. One of the other regulatory mechanisms is PIP₂ hydrolysis following the activation of Ca²⁺-dependent PLC, which also results in a switch of TRPM8 channels from a high to a low activity state. However, because there are multiple steps that are involved in PIP₂ hydrolysis following Ca²⁺ entry and also because PIP₂ on membranes is abundant, the consumption of PIP₂ is likely to be a slower process compared with CaM activation by Ca²⁺. Therefore, the switch of TRPM8 channels from a high to a low activity state following PIP₂ hydrolysis is also likely to be a slower process or tachyphylaxis at the macroscopic current level. TRPM8 channels may be dephosphorylated directly by protein phosphatase 1,2A or indirectly after PKC activation, which may lower the affinity between TRPM8 channels and PIP₂, thereby affecting the activity state of TRPM8 channels. DAG, diacylglycerol; IP₃, inositol 1,4,5-trisphosphate.

Different from acute desensitization, we found that tachyphylaxis was significantly attenuated by inhibiting PIP₂ hydrolysis or directly providing PIP₂. This result is consistent with the idea that PIP₂ is essential in maintaining TRPM8 at the high functional state and also identifies the specific role of PIP₂ hydrolysis in TRPM8 regulation. PIP₂ hydrolysis through Ca²⁺-dependent PLC can cause a gradual reduction of PIP₂ availability, which in turn switches the channel to the low functional state and thus tachyphylaxis (Fig. 9). At the single-channel level, the low functional state has low P_o with a predominance of long closed times. This is detectable on excision of the patches from cell-attached to inside-out configuration. PIP₂ prevents the channel from entering the low functional state and thereby prevents tachyphylaxis. Tachyphylaxis and the events that cause tachyphylaxis (e.g., PIP₂ hydrolysis) may start occurring along with the acute desensitization phase, as evidenced by the slow and mild current decay in the presence of CaM inhibitor ophiobolin A (Fig. 5D). However, tachyphylaxis is kinetically much slower compared with acute desensitization and thereby is less significant for the initial reduction of TRPM8 responses. The events that lead to tachyphylaxis appear to have a profound effect such that it is continuous even after TRPM8 activation is terminated. While we identify that PLC/PIP₂ is a main pathway leading to tachyphylaxis, we also identify that block of either PKC or PP1,2A reduced tachyphylaxis, a result that delineates the temporal phase of the effects mediated by PKC and PP1,2A in downregulation of TRPM8. A previous study suggested that PKC activation caused dephosphorylation rather than phosphorylation of TRPM8, which led to TRPM8 downregulation (Premkumar et al. 2005). It is possible that dephosphorylation of TRPM8 either directly by PP1,2A or indirectly after PKC activation may affect PIP₂ availability or the affinity of the TRPM8 channels for PIP₂, which could contribute to tachyphylaxis (Fig. 9). We showed that the inhibition of PKC or PP1,2A resulted in a reappearance of the acute desensitization in response to each repeated menthol application, a result similar to the experiments with PLC inhibitor or with the intracellular addition of PIP₂. This raises the possibility that PKC or PP1,2A activity may affect PIP₂ availability or affinity between PIP₂ and TRPM8 protein, thereby changing channel functional states.

In summary, CaM and PIP₂ inversely switch channel gating to regulate the functional states of TRPM8. Although both acute desensitization and tachyphylaxis are the results of the gating switches from a high functional state maintained by PIP₂ to a low functional state, the two processes are different, with the former requiring Ca²⁺-CaM and the latter dependent on PLC, PKC, and protein phosphatases. Thus, through multiple intracellular signaling pathways, TRPM8 functions in sensory neurons can be regulated in the forms of acute desensitization and tachyphylaxis.

GRANTS

This work was supported by National Institute of Dental and Craniofacial Research Grant DE-018661 to J. G. Gu.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).