Excitation and Modulation of TRPA1, TRPV1, and TRPM8 Channel-expressing Sensory Neurons by the Pruritogen Chloroquine

Background: Chloroquine (CQ) evokes the sensation of itch by exciting peripheral sensory neurons.

Results: CQ not only directly excites a diverse population of sensory neurons but also strongly modulates ion channels involved in pain and itch transduction.

Conclusion: CQ exerts widespread actions on peripheral sensory neurons.

Significance: Our results increase our understanding of the action of CQ on sensory neurons.

Abstract

The sensations of pain, itch, and cold often interact with each other. Pain inhibits itch, whereas cold inhibits both pain and itch. TRPV1 and TRPA1 channels transduce pain and itch, whereas TRPM8 transduces cold. The pruritogen chloroquine (CQ) was reported to excite TRPA1, leading to the sensation of itch. It is unclear how CQ excites and modulates TRPA1⁺, TRPV1⁺, and TRPM8⁺ neurons and thus affects the sensations of pain, itch, and cold. Here, we show that only 43% of CQ-excited dorsal root ganglion neurons expressed TRPA1; as expected, the responses of these neurons were completely prevented by the TRPA1 antagonist HC-030031. The remaining 57% of CQ-excited neurons did not express TRPA1, and excitation was not prevented by either a TRPA1 or TRPV1 antagonist but was prevented by the general transient receptor potential canonical (TRPC) channel blocker BTP2 and the selective TRPC3 inhibitor Pyr3. Furthermore, CQ caused potent sensitization of TRPV1 in 51.9% of TRPV1⁺ neurons and concomitant inhibition of TRPM8 in 48.8% of TRPM8⁺ dorsal root ganglion neurons. Sensitization of TRPV1 is caused mainly by activation of the phospholipase C-PKC pathway following activation of the CQ receptor MrgprA3. By contrast, inhibition of TRPM8 is caused by a direct action of activated Gα_q independent of the phospholipase C pathway. Our data suggest the involvement of the TRPC3 channel acting together with TRPA1 to mediate CQ-induced itch. CQ not only elicits itch by directly exciting itch-encoding neurons but also exerts previously unappreciated widespread actions on pain-, itch-, and cold-sensing neurons, leading to enhanced pain and itch.

Introduction

Pain is a distressing somatic sensation closely related to itch. Temperature-activated transient receptor potential (TRP)⁴ ion channels have important roles in the sensation of pain. For example, TRPV1 and TRPA1 channels mediate heat hyperalgesia and cold pain, respectively (1–5). By contrast, activation of the cold-activated TRPM8 channel by innocuous cooling inhibits pain (6, 7). Under pathological conditions such as inflammation, the activities of these ion channels are either sensitized (e.g. TRPV1 and TRPA1) or inhibited (e.g. TRPM8) through different signaling mechanisms, leading to a more painful outcome. Sensitization of TRPV1 and TRPA1 by inflammatory mediators that activate Gα_q-coupled G protein-coupled receptors (GPCRs) is caused mainly by activation of the phospholipase C (PLC) signaling pathway (8, 9). However, inhibition of TRPM8 by inflammatory mediators is largely independent of the PLC pathway; activated Gα_q protein instead directly inhibits TRPM8 (10).

In contrast to pain, the transduction of itch (pruritus) is less well-understood. There are two different types of itch, histamine-dependent (histaminergic) and histamine-independent (non-histaminergic), although the majority of pruritus is non-histaminergic and cannot be treated with antihistamines (11). Non-histaminergic itch is normally caused by exogenous pruritogens such as cowhage and chloroquine (CQ) (12–14).

Pruritogens cause itching by exciting sensory neurons largely through activation of Gα_q-coupled GPCRs, which then couple to downstream excitable ion channels, predominantly TRP ion channels. For example, histamine excites sensory neurons by activating TRPV1 through a signaling mechanism involving both the PLCβ3 and phospholipase A₂-lipoxygenase pathways (15–17). However, the non-histaminergic pruritogens serotonin and endothelin-1 induce itching without the involvement of TRPV1, although TRPV1-expressing neurons are required (16). On the other hand, CQ excites dorsal root ganglion (DRG) neurons and elicits itching by activating a Mas-related GPCR, MrgprA3, which then causes the opening of TRPA1 through a signaling mechanism involving Gβγ but not PLCβ (14, 16, 18). Therefore, different pruritogens seem to activate differential signaling mechanisms to excite DRG neurons.

Collectively, TRPV1 and TRPA1 are involved in the transduction of both pain and itch. Interestingly, activation of TRPM8 by cooling or by the cooling compound menthol inhibits pain and itch through both peripheral and central mechanisms (6, 19–21). It is thus not surprising that there is a complex and antagonistic relationship among the pain, itch, and cold pathways (20, 22–24). However, it is unclear how itching stimuli such as CQ affect the pain and cold pathways. In this study, we investigated how the non-histaminergic pruritogen CQ can excite and modulate the activity of TRPV1⁺, TRPA1⁺, and TRPM8⁺ DRG neurons, which are involved in transducing pain, itch, and cold.

EXPERIMENTAL PROCEDURES

Culture of DRG Neurons

DRG neurons were isolated from neonatal C57BL/6 mice as described previously (25). Briefly, mice of either sex were killed by cervical dislocation followed by decapitation. DRGs were collected from all cervical, thoracic, and lumbar segments and incubated in calcium- and magnesium-free Hanks' balanced salt solution (Invitrogen) containing 2.5 mg/ml type IV collagenase (Worthington Biochemicals) at 37 °C for 1 h. After incubation, DRGs were washed with DMEM (Invitrogen) followed by trituration with 25-gauge needles to dissociate neurons. The dissociated neurons were plated onto coverslips precoated with poly-l-lysine (Sigma) and laminin (BD Biosciences) and cultured in DMEM containing 10% fetal bovine serum (Invitrogen), 2 mm l-glutamine, 100 IU/ml penicillin, and 100 μg/ml streptomycin supplemented with 5 μm cytosine β-d-arabinofuranoside (Sigma) and 1× N2 supplement (Invitrogen). Cultured neurons were used for experiments within 24 h after plating. DRG neurons were also isolated from adult mice (22 days after birth) and cultured in a similar manner as the neonatal DRG neurons. NGF and Glial cell-Derived Neurotrophic Factor (GDNF) were not used for DRG cultures to avoid inducing the expression of TRP channels.

Culture and Transfection of Cell Lines

The HEK293 cell line and mouse embryonic fibroblast (MEF) cells lacking endogenous Gα_q/11 (26) were maintained as described previously (10). Briefly, cells were cultured in DMEM containing 10% fetal bovine serum, 2 mm l-glutamine, 100 IU/ml penicillin, and 100 μg/ml streptomycin. Cells were transfected with cDNAs encoding rat TRPM8 and MrgprA3 using PolyFect transfection reagent (Qiagen). Briefly, 4.0 μg of cDNAs was incubated with 20 μl of PolyFect reagent in 150 μl of serum-free DMEM for 10 min. After incubation, the solution was added to the cells and incubated for 24 h. Transfected cells were then replated onto small dishes for electrophysiological recording.

MEF cells were transfected using cell line nucleofection kits (Lonza). Briefly, cell pellets were resuspended in 100 μl of transfection solution containing 4.0 μg of cDNAs and electroporated using an Amaxa device. After transfection, cells were plated onto small dishes and used for electrophysiology recordings in 24 h.

Fluorescence Imaging

Calcium imaging was performed at room temperature as described previously (10, 25). Briefly, DRG neurons plated onto a coverslip were incubated with 7 μm Fluo-4-AM (Invitrogen) for 40 min at 37 °C. After loading, cell-containing coverslips were transferred to a custom-made chamber and continuously perfused with normal Hanks' solution containing 140 mm NaCl, 4 mm KCl, 10 mm HEPES, 1 mm MgCl₂, 1.8 mm CaCl₂, and 5 mm glucose (pH 7.4) with NaOH. Cell images were collected every 3 s using a Bio-Rad confocal microscope. Bradykinin (BK, 1 μm), histamine (100 μm), and CQ (1 mm) were pulsed onto cells for 2 min, whereas pulses of menthol (100 μm), allyl isothiocyanate (AITC) (100 μm), and capsaicin (500 nm) were applied for 15 s. All chemicals were applied every 4 min sequentially as appropriate to the individual experiment. To study the effect of PKC, cells were pretreated with the PKC inhibitor bisindolylmaleimide I (BIM; Calbiochem) as described previously (10). The transient receptor potential canonical (TRPC) inhibitor YM58483 (BTP2; Tocris Bioscience), the TRPC3 inhibitor Pyr3 (Tocris Bioscience), the TRPA1 inhibitor HC-030031 (Tocris Bioscience), and ruthenium red (Sigma) were applied as indicated in individual experiments. Ionomycin was applied at the end of experiment to allow maximal calcium influx. Neurons unresponsive to ionomycin were excluded for further analysis. We considered that a chemical induced a calcium response in DRG neurons if the difference between the elicited peak calcium response and the base-line fluorescence (ΔF/F_max) was >0.06. We also performed similar calcium imaging experiments in calcium-free Hanks' solution. Under these conditions, no significant calcium responses (0 of 694 neurons) were observed for all chemicals, excluding the possibility of the contribution of intracellular calcium stores to excited calcium responses.

To study the modulation of TRPV1 and TRPM8 by CQ, the effect of CQ was quantified as a response ratio by dividing the fifth peak response amplitude by the fourth peak response amplitude. In control experiments on cells not exposed to CQ, the distribution of response ratios was found to be well fitted by a normal distribution (see Fig. 3E), from which a threshold ratio was derived at a 95% confidence level and used to determine cells significantly sensitized by CQ.

FIGURE 3.

CQ sensitizes TRPV1 in DRG neurons. A and B, the calcium response elicited by capsaicin (Cap; 100 nm) was enhanced by CQ (1 mm) in a DRG neuron; however, the neuron did not respond to 100 μm AITC, and CQ did not elicit a calcium increase irrespective of whether CQ was applied before (B) or after (A) capsaicin. Men, menthol. C and D, representative calcium response traces from a DRG neuron in an experiment similar to that in A, but CQ induced a transient calcium increase. Note that the application sequence of AITC in D was different from that in C. AITC was added after the second application of CQ in D. E, distribution of TRPV1-dependent calcium response ratios after (fifth response) and before (fourth response) CQ in DRG neurons from experiments similar to those in A. The number of cells was 137 for the control cells and 581 for the CQ-treated cells. F, summary of experiments similar to those in E. The number of experiments was 5 for the control (Con) cells, 11 for the 1 mm CQ-treated cells, and 7 for the BIM-pretreated cells (a total of 108 cells). Data are means ± S.E. ***, p < 0.001 compared with the control; ###, p < 0.001 compared with the second bar.

Tubby-R332H-cYFP (where “c” indicates C-terminal; kindly provided by Dr. Gerald Hammond) translocation was determined by live scanning using a Leica confocal microscopy. Images of MEF cells transfected with the fluorescence probe and MrgprA3 were collected every 0.75 s. The translocation of Tubby-R332H-cYFP was quantified by calculating the ratio of membrane fluorescence to cytosol fluorescence using ImageJ software.

Single-cell RT-PCR

Neurons were first examined by calcium imaging. CQ-sensitive neurons in each category were then aspirated into a glass electrode and transferred into a PCR tube. The total RNA was isolated from the collected neurons using the RNeasy micro kit (Qiagen) followed by reverse transcription with Superscript II reverse transcriptase (Invitrogen) according to the manufacturer's instructions. The products of reverse transcription were then amplified by PCR with the primers listed in Table 1. Cycling conditions were as follows: 2 min at 95 °C; 50 cycles of 50 s at 95 °C, 45 s at 60 °C, and 46 s at 72 °C; and 5 min at 72 °C. The PCR products were then analyzed on 1% agarose gel.

TABLE 1.

Primer sequences for single-cell RT-PCR

Gene	Primer sequences	Product length	GenBankTM accession no.
		bp
MrgprA3		150	NM_153067.2
Forward	5′-CGACAATGACACCCACAACAA-3′
Reverse	5′-GGAAGCCAAGGAGCCAGAAC-3′
TRPV1		362	NM_001001445.1
Forward	5′-CTGGAGCTGTTCAAGTTCACC-3′
Reverse	5′-TTGGTGTTCCAGGTAGTCCAG-3′
TRPA1		684	AY231177
Forward	5′-CCAAGATGCCTTCAGCACCC-3′
Reverse	5′-GGGTGGCTAATAGAACAATGTGTTTTAGTC-3′
TRPM8		458	NM_134252.
Forward	5′-GTGGGAGCAACTGTCTGGAGC-3′
Reverse	5′-GTGTCCATAACGTTCCATAGGTCG-3′
TRPC3		317	NM_019510.2
Forward	5′-GGAACTGGGCATGGGTAACTC-3′
Reverse	5′-CACTGGGGTTCAGTTTCTCACTG-3′
GAPDH		302	NM_008084
Forward	5′-CATCCATGACAACTTTGGCA-3′
Reverse	5′-CCTGCTTCACCACCTTCTTG-3′

Electrophysiology

Whole-cell patch recordings were performed largely as described previously (9, 10). Briefly, patch electrodes were filled with an internal solution of 140 mm KCl, 2.0 mm MgCl₂, 5.0 mm EGTA, and 10 mm HEPES (pH 7.4) with KOH. Standard Hanks' solution was used as the external solution. TRPV1 inward currents were recorded at a holding potential of −60 mV. For recordings of the TRPM8 current, we used calcium-free solution in which 1.8 mm CaCl₂ was replaced with 5 mm EGTA to prevent the interference of calcium-dependent desensitization. TRPM8 inward and outward currents were measured at holding potentials of −60 and +60 mV, respectively. Cells were pretreated with 1 mm CQ for 1 min before break in to the whole cell to measure TRPM8 currents activated by menthol. Series resistance was routinely ≤10.0 megohms and was compensated by >70%. All recordings were made at room temperature (24 °C) with an Axopatch 200B patch clamp amplifier (Axon) in conjunction with pClampex version 10.2 software (Molecular Devices). Signals were analog-filtered using a 1-kHz low-pass Bessel filter.

Co-immunoprecipitation

This was performed as described previously with some modifications (10). Briefly, HEK293 cells expressing TRPM8-V5 and MrgprA3 were solubilized, and TRPM8 was precipitated using anti-V5 antibody (Invitrogen) and protein A-agarose (Santa Cruz Biotechnology). The coprecipitates were then analyzed on 10% SDS-polyacrylamide gel and transferred to a blot. Gα_q protein was detected with anti-Gα_q polyclonal antibody (Santa Cruz Biotechnology). The V5 tag does not affect the binding of Gα_q to TRPM8 because Gα_q can also be similarly precipitated using anti-TRPM8 polyclonal antibody (10).

Statistics

All data are means ± S.E. Differences between groups were assessed by one-way analysis of variance with the Bonferroni post hoc test. Results were considered significant at p < 0.05.

RESULTS

CQ Excites a Subpopulation of Sensory Neurons Expressing TRPV1, TRPA1, and TRPM8

CQ is traditionally used to treat malaria. One of the common side effects associated with CQ therapy is the production of serious itching. CQ-induced itching has recently been found to be caused by the excitation of the TRPA1 ion channel in DRG neurons after activation of the CQ receptor MrgprA3 (18, 14). However, it is unclear which populations of DRG neurons are excited by CQ and how they relate to TRPV1⁺, TRPA1⁺, and TRPM8⁺ neurons, which mediate the sensations of pain, itch, and cold.

To investigate this question, we assessed the responses of neonatal DRG neurons after sequential exposure to CQ, the TRPM8 agonist menthol, the TRPA1 agonist AITC, and the TRPV1 agonist capsaicin by monitoring intracellular calcium rises. Two minutes of CQ treatment rapidly induced calcium increases in a subpopulation of neurons (Fig. 1, A–D). About 12.8% (344 of 2693 total neurons) of DRG neurons were found to be excited by CQ (Fig. 1, F and H). Strikingly, only 43.3% (149 of 344) of these neurons also responded to AITC and thus coexpressed TRPA1 (Fig. 1B, F, and H); some of these neurons (99 of 149) responded to both AITC and capsaicin and thus coexpressed both TRPA1 and TRPV1 (Fig. 1, D and F). Notably, 49.7% (171 of 344) of CQ-responsive neurons did not coexpress TRPA1 but expressed solely TRPV1, and the remaining 7% (24 of 344) of CQ-sensitive neurons expressed neither TRPV1 nor TRPA1 (Fig. 1, C, F, and H). Consistent with the pharmacological data, the single-cell RT-PCR revealed that MrgprA3 mRNA was expressed in all CQ-responsive neurons (Fig. 1E). Some of them coexpressed TRPA1 and/or TRPV1 mRNA, whereas others coexpressed neither TRPV1 nor TRPA1 mRNA (Fig. 1E). Fig. 1F shows a summary of percentages of the individual populations of CQ-excited neurons described above related to a total of 2693 neurons isolated from neonatal mice. Similar percentages of CQ-excited DRG neurons for these individual populations were also obtained from a total of 1450 neurons isolated from adult mice (Fig. 1G). Hence, the properties of CQ-excited neurons do not change significantly after postnatal development.

FIGURE 1.

Direct excitation of DRG neurons by CQ. A, fluorescence images of DRG neurons loaded with Fluo-4-AM under basal conditions and after stimulation with CQ (1 mm, 2 min) and AITC (100 μm, 15 s). The fourth panel is a bright-field image of the neurons. Arrows indicate neurons responding to CQ. Scale bars = 10 μm. B–D, representative traces from DRG neurons as described for A activated sequentially by CQ (1 mm), menthol (Men; 100 μm), AITC (100 μm), and capsaicin (Cap; 500 nm). Ionomycin (Iono; 10 μm) was added at the end. CQ-responsive neurons either coexpressed TRPA1 (B) or TRPV1 (C) or both TRPV1 and TRPA1 (D). E, expression of TRP channel mRNA in individual DRG neurons amplified by RT-PCR from different subpopulations categorized by calcium imaging as shown in A–D. +, neurons sensitive to CQ, menthol, AITC, or capsaicin; −, insensitive neurons. The fifth lane is the amplification from the control bath solution. F and G, summary of percentages of CQ-responsive, capsaicin-sensitive, and AITC-responsive neurons related to a total of 2693 neurons isolated from neonatal mice (F) or 1450 neurons isolated from adult mice (G) in experiments similar to those in A–C. H, the Venn diagram illustrates the mutual relationship of TRPV1⁺, TRPA1⁺, and TRPM8⁺ (M8) neurons with CQ-responsive neurons. The number of neurons is shown for each population and was obtained from a total of 2693 neonatal neurons.

We also found that a small proportion (14.2%, 49 of 344) of CQ-excited neurons was excited by menthol and thus coexpressed TRPM8 (Fig. 1H). Most of these neurons (76%) also coexpressed TRPV1 and/or TRPA1 (Fig. 1H) and were thus contained within TRPV1⁺ and TRPA1⁺ populations. Taken together, CQ excited three main populations of neurons coexpressing either TRPA1 or TRPV1 or neither channel.

CQ Excites Sensory Neurons through Both TRPA1-dependent and TRPC3-dependent Mechanisms

To examine whether coexpressed TRPA1 or TRPV1 in DRG neurons is a bona fide ion channel responsible for mediating responses caused by CQ, we treated CQ-responsive neurons with the specific TRPA1 blocker HC-030031, the TRPV1 blocker capsazepine, or the general TRP channel blocker ruthenium red. We found that CQ-excited responses were completely blocked by the specific TRPA1 antagonist HC-030031 in all neurons that were also responsive to AITC (Fig. 2, A and F). This experiment shows that TRPA1 is the sole molecule mediating the excitation of this population of DRG neurons evoked by CQ, consistent with the proposed role of TRPA1 in CQ-induced itch (18). However, HC-030031 had no effect on the responses excited by CQ in neurons that were unresponsive to AITC (Fig. 2, B and F). Because most of this population of neurons coexpressed TRPV1, we then investigated whether TRPV1 might mediate the excitation of this population of neurons by treating neurons with the specific TRPV1 antagonist capsazepine. Fig. 2F shows that capsazepine failed to prevent CQ-induced responses in neurons that were responsive to the TRPV1 agonist capsaicin, suggesting that ion channels other than TRPV1 mediate the excitation of TRPV1⁺ neurons elicited by CQ. Calcium responses evoked by CQ are not likely to be caused by the release of intracellular calcium store because no calcium responses were observed when similar experiments were performed in a calcium-free solution (Fig. 2C). The CQ-induced calcium responses were prevented, however, by the general TRP channel blocker ruthenium red in all neurons, whether AITC-sensitive or not (Fig. 2, D and F), further supporting the notion that the calcium responses evoked by CQ are caused by extracellular calcium influx through Ca²⁺-permeable TRP ion channels in the plasma membrane (14). These results demonstrate that CQ-induced excitation of sensory neurons is mediated by TRPA1 only in neurons coexpressing TRPA1 or TRPA1/TRPV1, which account for 43.3% of the total CQ-excited neurons. The responses of the remaining populations of neurons (56.7%) to CQ are not mediated by either TRPA1 or TRPV1 but instead by an as-yet-unidentified TRP ion channel.

FIGURE 2.

CQ excites DRG neurons through both TRPA1- and TRPC3-dependent mechanisms. A and B, representative calcium responses from three DRG neurons activated by CQ (1 mm) and AITC (100 μm). The TRPA1 antagonist HC-030031 (10 μm) was added together with the second CQ application as indicated. Iono, ionomycin. C–E, typical calcium responses from DRG neurons excited by CQ followed by exposure to AITC (100 μm) and capsaicin (Cap; 500 nm). The calcium-free solution, ruthenium red (RR; 10 μm), and BTP2 (10 μm) were perfused together with the second CQ addition as indicated. F, summary of ratio percentages between the second CQ-excited response and the first CQ-induced response obtained from neonatal DRG neurons in experiments similar to those in A–E. +, CQ-sensitive, AITC-sensitive, and capsaicin-sensitive neurons; −, insensitive neurons. The TRPA1 antagonist HC-030031 (10 μm), the TRPC channel blocker BTP2 (10 μm), the TRPC3 channel inhibitor Pyr3 (10 μm), the TRPV1 antagonist capsazepine (CSZ; 10 μm), and ruthenium red were added as indicated. The numbers of neurons are given above each bar. Data are means ± S.E. ***, p < 0.001; NS, not significant. All results are compared with the first bar.

TRPC channels are Ca²⁺-permeable nonselective cation channels. Several subtypes are abundantly expressed in DRG neurons (27). To determine the role of TRPC channels in TRPA1-independent excitation of DRG neurons caused by CQ, DRG neurons were treated with the general TRPC channel inhibitor BTP2. The CQ-induced responses in TRPA1⁻ neurons (insensitive to AITC) were greatly reduced when neurons were pretreated with BTP2 (Fig. 2, E and F). TRPC3 has recently been found to be involved in the excitation of sensory neurons induced by the IgG immune complex (28). We then examined the possible contribution of TRPC3 to CQ-excited responses by treating neurons with the selective TRPC3 inhibitor Pyr3. Fig. 2F shows that Pyr3 prevented CQ-elicited responses in TRPA1⁻ neurons to a similar extent as BTP2. The single-cell RT-PCR further revealed that this population of DRG neurons coexpressed TRPC3 mRNA (Fig. 1E). These data demonstrate that the excitation of DRG neurons in TRPA1⁻ DRG neurons caused by CQ is mediated by TRPC3.

CQ Sensitizes TRPV1 in DRG Neurons

Because most of the CQ-excited neurons (78.5%) coexpressed TRPV1 (Fig. 1H), and yet TRPV1 is not a mediator for the direct excitation of TRPV1⁺ neurons by CQ, we wondered whether CQ might be involved in the modulation of TRPV1 function instead, thereby affecting the sensations of pain and itch. To test this hypothesis, we monitored TRPV1-mediated calcium responses evoked by capsaicin in DRG neurons. A short treatment (2 min) with CQ caused a robust enhancement of TRPV1-mediated calcium responses (Fig. 3A). However, CQ treatment itself did not generate calcium signals. To exclude the possible effect of repeated capsaicin stimulation on the CQ-excited calcium response, neurons were also stimulated first with CQ and then with capsaicin. The second pulse of CQ similarly enhanced the TRPV1-mediated calcium increase despite the fact that CQ did not evoke a calcium response by itself (Fig. 3B). In fact, 93% of the neurons that were sensitized by CQ did not show a CQ-excited response. TRPV1 is thus unlikely to be a downstream effector ion channel underlying CQ-initiated excitation of neurons, in agreement with our findings (above) and the previous observations of others (18). The absence of CQ-excited calcium responses in these sensitized neurons may be caused by the lack of a downstream effector ion channel such as TRPA1. Indeed, most of these neurons were not co-excited by AITC (Fig. 3, A and B); however, in some neurons (7%), CQ not only sensitized TRPV1 but also excited a calcium response itself despite the lack of expression of TRPA1 (Fig. 3C). Surprisingly, CQ did not cause the sensitization of TRPV1 in some neurons (25 of 581 TRPV1⁺ neurons) that were also excited by CQ (Fig. 3D). Presumably, CQ excites these neurons through an MrgprA3-independent mechanism.

Interestingly, although CQ excited only 24.5% (270 of 1100) of TRPV1⁺ neurons (Fig. 1H), CQ caused sensitization in 51.9% of TRPV1⁺ neurons (Fig. 3, E and F), comparable with the sensitization induced by BK (9). These data suggest that the CQ receptor is expressed in at least twice as many sensory neurons as what we estimated above by monitoring CQ-excited calcium responses and that the CQ receptor is widely distributed across diverse populations of DRG neurons and thus may have much more widespread actions on DRG neurons than previously appreciated.

CQ-induced Sensitization of TRPV1 Involves the PLC-PKC Pathway

CQ acts on the CQ receptor MrgprA3, which belongs to a family of orphan GPCRs known as Mrgs (Mas-related genes). Two other members, MrgA1 and MrgC11, are coupled to the Gα_q/11 pathway (29). To test whether MrgprA3 is also coupled to Gα_q/11, we used Tubby-R332H-cYFP, a sensitive reporter for the membrane lipid phosphatidylinositol 4,5-bisphosphate (PIP₂), to monitor whether activation of MrgprA3 causes PIP₂ depletion. Fig. 4A shows that CQ stimulated a rapid translocation of Tubby fluorescence away from the plasma membrane, indicating depletion of PIP₂. Moreover, CQ caused Tubby translocation with similar kinetics and to a similar extent as BK (10). These experiments demonstrate that MrgprA3 is a Gα_q/11-coupled receptor and that activation of MrgprA3 has a similar ability to activate PLCβ and hydrolyze PIP₂ as activation of the BK receptor B2R.

FIGURE 4.

The PLC-PKC pathway is involved in the sensitization of TRPV1 induced by CQ. A, translocation of Tubby-R332H-cYFP caused by CQ in HEK293 cells expressing MrgprA3. CQ (1 mm) was added at 70 s. Scale bars = 10 μm. B, quantification of relative membrane Tubby fluorescence signal as a function of time in A. This experiment was repeated at least four times with similar results. PM, plasma membrane. C, CQ-sensitized TRPV1 inward current activated by 50 nm capsaicin (Cap; 5 s) recorded from a HEK293 cell expressing MrgprA3 and TRPV1. The dotted line represents zero current. D, CQ did not cause sensitization of the TRPV1 inward current recorded from a HEK293 cell expressing MrgprA3, TRPV1, and PLCβ-ct. The dotted line represents zero current. E, summary of results similar to those in C and D. Sensitization of TRPV1 was blocked by 2.5 μm U73122, by coexpression with PLCβ-ct, by 1 μm BIM, or by double-mutant (DM) TRPV1 S502A/S801A. All error bars are means ± S.E. The number of experiments is given above each bar. ***, p < 0.001 compared with the control (Con); ## p < 0.01; ### p < 0.001 compared with the second bar.

CQ excites the TRPA1 channel through Gβγ, which is released after activation of the CQ receptor MrgprA3 (18). We were therefore interested in investigating the underlying signaling mechanisms for CQ-induced sensitization of TRPV1. We recorded TRPV1 currents in HEK293 cells expressing both TRPV1 and MrgprA3. Consistent with CQ-induced sensitization of TRPV1 in DRG neurons, CQ potently sensitized the TRPV1 inward current to a similar extent as BK (Fig. 4C) (9). The sensitization of TRPV1 by CQ was completely inhibited by the PLC inhibitor U73122 and also completely abolished by coexpression of the C-terminal domain of PLCβ (PLCβ-ct) (Fig. 4D), which is used to sequester endogenous Gα_q and thus block the PLC pathway (30). CQ-induced sensitization of TRPV1 was also prevented by the specific PKC inhibitor BIM in both DRG neurons and the exogenous expression system (Figs. 3F and 4E). Mutation of two PKC phosphorylation sites on TRPV1 similarly prevented the sensitization of TRPV1 by CQ (Fig. 4E). These experiments suggest that the PLC-PKC pathway plays a key role in the sensitization of TRPV1 caused by CQ.

CQ Inhibits TRPM8 in a Subpopulation of DRG Neurons

In contrast to TRPV1 and TRPA1, which mediate the sensations of pain and itch, activation of TRPM8 by moderate cooling or menthol inhibits pain and itch (20). Because CQ-excited neurons were also found to overlap partly with TRPM8⁺ neurons (Fig. 1H), we were interested in determining whether CQ could also modulate TRPM8 function in DRG neurons. To this end, we monitored TRPM8-mediated calcium responses elicited by menthol. Importantly, menthol was also reported to activate the TRPA1 channel (31). To identify TRPM8⁺ neurons specifically, we therefore also applied AITC and selected neurons that were both menthol-sensitive and AITC-insensitive. In fact, only 3.2% of AITC-sensitive neurons were also menthol-sensitive (Fig. 1H), suggesting that menthol largely activates TRPM8⁺ neurons. Fig. 5A shows that 2 min of exposure to CQ significantly reduced the TRPM8-mediated calcium increase and that CQ itself also induced a transient calcium response. In some neurons, CQ still inhibited TRPM8 even if CQ did not elicit a calcium response (Fig. 5B); these neurons probably expressed the CQ receptor MrgprA3 but lacked downstream effector ion channels for excitation. Overall, CQ inhibited 48.8% (21 of 43) of TRPM8⁺ neurons. The mean ratio of calcium responses after and before CQ treatment was significantly reduced (Fig. 5, C and D).

FIGURE 5.

CQ inhibits TRPM8 in DRG neurons. A and B, CQ-inhibited calcium increase elicited by the TRPM8 agonist menthol (Men; 100 μm) in a DRG neuron. 100 μm AITC and capsaicin (Cap; 500 nm) were added as indicated. C, histogram distribution of calcium response ratios after (fifth response) and before (fourth response) CQ treatment. The number of cells was 30 for the control cells and 43 for the CQ-treated cells. D, summary of results similar to those in A after pretreatment with 1 μm BIM (38 cells). All data are means ± S.E. ***, p < 0.001 compared with the first bar. Con, control. E, summary of TRPM8-mediated calcium response ratios between after and before BK or CQ treatment in experiments similar to those in A and D and those obtained previously (10), but only inhibited neurons were included. The numbers of inhibited neurons are given above each bar. All data are means ± S.E. ***, p < 0.001 compared with the first bar.

We have previously found that TRPM8 can also be inhibited by BK in DRG neurons (10). Compared with BK-elicited inhibition of TRPM8, the TRPM8-mediated calcium response ratio after and before CQ treatment was significantly larger than that caused by BK (Fig. 5E), showing that CQ inhibits TRPM8 to a lesser extent than BK.

CQ-mediated Inhibition of TRPM8 Involves Direct Actions of Activated Gα_q/11 on TRPM8

We then investigated the mechanisms underlying CQ-elicited inhibition of TRPM8. In contrast to PKC-mediated sensitization of TRPV1 caused by CQ, CQ-induced inhibition of TRPM8 was not prevented by treatment with the specific PKC inhibitor BIM (Fig. 5D), suggesting that inhibition of TRPM8 by CQ is not caused by PKC.

To further investigate possible signaling mechanisms, TRPM8 inward and outward currents were recorded in HEK293 cells expressing both TRPM8 and MrgprA3. Fig. 6A shows that CQ stimulation inhibited TRPM8 inward currents but not outward currents. It is noteworthy that although MrgprA3 and the BK receptor B2R are both Gα_q-coupled receptors and can initiate PIP₂ depletion to a similar extent (Fig. 4, A and B) (10), CQ caused TRPM8 inhibition to a much lesser extent than BK (Fig. 6B) (10), consistent with the above findings in DRG neurons. CQ-induced inhibition of TRPM8 was not reversed by either the PLC inhibitor U73122 or coexpression of PLCβ-ct (Fig. 6, A and B), both of which completely inhibited CQ-induced sensitization of TRPV1 (Fig. 4E). These data suggest that inhibition of TRPM8 by CQ takes place through a mechanism independent of the PLC-PKC pathway.

FIGURE 6.

CQ-induced TRPM8 inhibition is caused by direct action of activated Gα_q. A, inward and outward currents (at −60 and +60 mV) activated by menthol (200 μm, 5 s) in HEK293 cells expressing TRPM8 and MrgprA3 were inhibited by pretreatment with CQ (1 mm, 1 min) applied alone or together with U73122 (2.5 μm) or after coexpression with PLCβ-ct. The dotted line represents zero current. Con, control. B, summary of peak currents in experiments similar to those in A. The number of experiments is shown above each bar. C, representative traces of TRPM8 inward and outward currents (at −60 and +60 mV) activated by menthol (200 μm, 5 s) in Gα_q/11^−/− MEF cells transfected with MrgprA3 and TRPM8. Currents are shown before and after CQ (1 mm, 1 min) treatment. The dotted line represents zero current. D, summary of results similar to those in C. The number of experiments is shown above each bar. All data are means ± S.E. **, p < 0.01; NS, not significant. E, HEK293 cells expressing TRPM8-V5 and MrgprA3 were stimulated with 1 mm CQ for 5 min. TRPM8 was then immunoprecipitated (IP) with anti-V5 antibody, and coprecipitated Gα_q was detected by anti-Gα_q antibody The same blot was stripped and redetected with anti-V5 antibody. TCL, total cell lysate; IB, immunoblot. Molecular masses are shown on the right.

We have previously found that BK and histamine inhibit TRPM8 via a direct action of activated Gα_q on TRPM8 (10). To test whether this mechanism also applies to CQ-induced inhibition of TRPM8, we performed similar experiments in MEF cells lacking endogenous Gα_q/11. As expected, inhibition of TRPM8 was absent after CQ treatment in MEF cells (Fig. 6C). However, coexpression of the 3Gα_qiq chimera, which is unable to couple to PLCβ (10), rescued inhibition of TRPM8 caused by CQ (Fig. 6, C and D). Furthermore, TRPM8 interacted with Gα_q as indicated by the coprecipitation of Gα_q with TRPM8 (Fig. 6E). However, CQ did not induce significant additional binding of Gα_q to TRPM8 (Fig. 6E), suggesting that a conformation change in activated Gα_q is sufficient to inhibit TRPM8. These experiments indicate that inhibition of TRPM8 evoked by CQ is caused by a direct action of activated Gα_q on TRPM8 independent of downstream signaling pathways.

DISCUSSION

CQ causes itching by activating the CQ receptor MrgprA3, which then couples to excitation of the TRPA1 channel (14, 18). Hence, both MrgprA3 and TRPA1 are essential for mediating CQ-induced itch. In this study, we found that CQ excited a diverse population of DRG neurons and that TRPA1 mediated responses in only 43.3% of CQ-excited DRG neurons. The responses in the remaining 56.7% of CQ-excited neurons were not mediated by either TRPA1 or TRPV1 despite the fact that CQ-excited neurons (78.5%) coexpressed TRPV1. All CQ-excited responses were abolished, however, by the general TRP channel blocker ruthenium red. The CQ-induced responses in a subpopulation of TRPA1⁻ DRG neurons were also inhibited by both a general TRPC channel blocker and a selective TRPC3 inhibitor. These results suggest that the TRPC3 channel may also be involved in CQ-induced itch.

Very recently, MrgprA3⁺ neurons were reported to be itch-specific DRG neurons dedicated to the transduction of itch, and it was shown that activation of TRPV1 in MrgprA3⁺ neurons caused itch but not pain behavior (32). TRPV1 was also reported to mediate histamine-induced itch (15). Interestingly, we found that TRPV1 was robustly sensitized by CQ, although TRPV1 did not mediate CQ-induced excitation of DRG neurons directly. This finding suggests that indirect sensitization of TRPV1 within MrgprA3⁺ itch-specific neurons could lead to the sensitization of both histaminergic and non-histaminergic itching responses. Hence, CQ not only causes itching through direct excitation of itch-transducing neurons via TRPA1 but also enhances itching through facilitation of activation of TRPV1.

Additionally, we found that activation of MrgprA3 by CQ inhibited the cold-sensitive ion channel TRPM8. Activation of TRPM8 by cooling was reported to inhibit pain and itch (20); thus, concomitant inhibition of TRPM8 could also lead indirectly to enhanced sensations of pain and itch.

CQ acts on multiple TRP ion channels; it excites TRPA1, sensitizes TRPV1, and inhibits TRPM8. Strikingly, CQ produces these actions through completely different signaling mechanisms. The CQ receptor MrgprA3 is a Gα_q-coupled GPCR. Activation of TRPA1 by CQ is believed to be mediated by Gβγ released after activation of MrgprA3. In contrast, sensitization of TRPV1 elicited by CQ is mediated mainly by activation of the PLC-PKC pathway. However, CQ-evoked inhibition of TRPM8 is independent of either of these two pathways; instead, a direct action of activated Gα_q on the TRPM8 channel seems to be responsible (Fig. 7).

FIGURE 7.

CQ excites TRPA1 and TRPC3, sensitizes TRPV1, and inhibits TRPM8 through different signaling mechanisms. Binding of CQ to the CQ receptor MrgprA3 causes activation of Gα_q and release of Gβγ. The released Gβγ then activates the TRPA1 channel, leading to opening of the channel and calcium influx. Meanwhile, activated Gα_q activates PLCβ, leading to excitation of the TRPC3 channel and calcium influx. Activated Gα_q also stimulates PKC and increases phosphorylation of TRPV1, leading to the sensitization of TRPV1 and subsequently enhanced itch. Activated Gα_q also directly inhibits TRPM8 by binding to TRPM8 without the need to involve the PLC pathway. Inhibition of TRPM8 probably also contributes indirectly to enhanced pain and itch.

Both the CQ receptor MrgprA3 and the BK receptor B2R are Gα_q/11-coupled receptors. Activation of these two receptors depleted PIP₂ to a similar extent as shown in Fig. 4 (A and B) and in a similar experiment with BK reported previously (10). Furthermore, CQ caused sensitization of TRPV1 to a similar extent as BK in both DRG neurons and transfected cells (Fig. 4E) (9), suggesting that a similar degree of activation of downstream PKC has been triggered by CQ and BK. Despite these similarities between CQ- and BK-initiated signaling, CQ inhibited TRPM8 to a much lesser extent than BK, suggesting that PIP₂ cleavage caused by the activation of PLCβ may not be the major mediator of the modulation of TRPM8 by activated Gα_q/11. We have also recently found that activated Gα₁₁ inhibits TRPM8 to a much smaller extent than activated Gα_q, although these proteins are indistinguishable in activating PLC and hydrolyzing PIP₂ (33). The weaker inhibition of TRPM8 caused by CQ could thus be due to differential expression of Gα₁₁ in MrgprA3⁺ neurons. It could also be caused by the differential binding of Gα_q/11 to TRPM8 in MrgprA3⁺ neurons, leading to different levels of TRPM8 inhibition.

Taken together, our data suggest that CQ produces a much broader action than previously appreciated by acting on multiple TRP ion channels, including TRPV1, TRPA1, TRPC3, and TRPM8, across diverse populations of DRG neurons, leading to enhanced pain and itch. TRP ion channels thus represent attractive targets for the treatment of both pain and itch.