The Quaternary Lidocaine Derivative, QX-314, Exerts Biphasic Effects on Transient Receptor Potential Vanilloid Subtype 1 Channels In Vitro 

• The charged lidocaine derivative QX-314 interacts with transient receptor potential vanilloid (TRPV) channels to enter sensory fibers and produce analgesia

• Whether QX-314 directly activates or inhibits these channels is not known

• In injected oocytes, QX-314 inhibited TRPV1 channels at micromolar and activated TRPV1 channels at millimolar concentrations, without affecting TRPV4 channels.

THE transient receptor potential vanilloid subfamily member 1 (TRPV1) channel or “capsaicin receptor” recently has garnered interest as a potential target for novel analgesic compounds.1–6Among the investigated candidate molecules is the quaternary lidocaine derivative, QX-314 (N -[2,6-dimethylphenylcarbamoylmethyl]triethylammonium), which features an additional N -ethyl group attached to the amine moiety. As a result, the traditional view has been that QX-314 is membrane-impermeable and blocks Na⁺-dependent action potentials only when administered intracellularly, rendering the agent devoid of clinically useful local anesthetic (LA) activity. However, we recently found that QX-314 concentration-dependently produces reversible robust long-lasting nociceptive, sensory, and motor blockade with a slow onset in animal models in vivo .7A subsequent report raised the intriguing possibility that QX-314 can be selectively navigated into nociceptive neurons through the pore of activated TRPV1 channels, providing a potential mechanism for achieving nociceptive-specific blockade.8Additional pharmacologic exploration of this possibility confirmed that coapplication of QX-314 with the TRPV1 agonist, capsaicin, significantly accelerates the onset of sensory blockade attributable to QX-314.9However, direct mechanistic insight into the interaction of QX-314 with TRPV1 has been lacking.

TRPV1 is the best-studied member of the transient receptor potential group of ion channels, a ubiquitously expressed and diverse class of environmental-sensing proteins. Structurally, TRPV1 is a homotetramer, with each monomer containing six transmembrane-spanning α helical domains and a reentrant loop between transmembrane domains 5 and 6 that contributes to the pore region.10In terms of its function, TRPV1 is a nonspecific cation channel with a preference for Ca²⁺that exhibits pronounced expression in thin myelinated and unmyelinated axons (AΔ and C fibers, respectively), as well as somata in sensory (dorsal root, trigeminal, and nodose) ganglia.11The peripheral terminals of capsaicin-sensitive neurons are sites of release for various proinflammatory neuropeptides, such as substance P and calcitonin gene-related peptide, which in turn initiate the biochemical cascade that underlies neurogenic inflammation.12The central fibers of capsaicin-sensitive neurons enter the dorsal horn of the spinal cord, where they form synapses with second-order neurons.11As such, TRPV1 has been well established as an important integrator of nociceptive information.10,13TRPV1 is unique in that it is a polymodal nociceptor exhibiting a dynamic threshold of activation10: agents involved in inflammation act together to lower the activation threshold.13,14The list of agents that can activate and/or sensitize TRPV1 is extensive and in addition to acidic and basic pH,14,15includes bradykinin,16,17nerve growth factor,17anandamide,18arachidonic acid metabolites,19lipoxygenase products,20leukotriene B,21prostaglandins,22adenosine triphosphate,23jelly fish and spider venoms,24,25volatile anesthetics,26and propofol (providing at least a partial explanation for its injection pain).27 

In addition, recent research has pointed to a direct interaction between TRPV1 channels and amino amide LAs. For example, lidocaine activates TRPV1 and induces a TRPV1-dependent release of calcitonin gene-related peptide, which has been suggested as a mechanism for some of its adverse and toxic effects.28We recently found that intrathecal administration of QX-314 at concentrations as low as 0.5 mM produced unacceptable adverse effects in mice (including death at ≥ 5 mM).29These effects occurred at lower concentrations than those associated with robust motor blockade, suggesting a low therapeutic index for use in spinal anesthesia. The findings also suggested a distinct neurotoxic action of QX-314 compared with lidocaine29: whereas lidocaine neurotoxicity mainly manifests as irreversible conduction blockade, intrathecal QX-314 acutely produced marked irritable, nocifensive-type behavior in the animals. Although QX-314 holds promise as an LA to produce long-lasting peripheral antinociception, it is imperative on the basis of these observations that more preclinical studies be conducted to better define the mechanisms underlying the therapeutic and adverse effects of QX-314 before it is tested in humans.

Here we considered the possibility that QX-314, at millimolar concentrations, may produce direct TRPV1 activation as a mechanism for acute irritation. To test this hypothesis directly, we conducted an in vitro  study to examine the mechanistic basis by which QX-314 acts on TRPV1 channels. Our results demonstrate that QX-314 exerts biphasic effects on expressed TRPV1 channels, where micromolar application potently inhibits capsaicin-induced cation currents, whereas millimolar concentrations produce robust and reproducible channel activation.

All drugs were obtained from Sigma-Aldrich Canada (Oakville, Ontario, Canada). Capsaicin (8-methyl-N -vanillyl-6-nonenamide) was dissolved in absolute ethanol to produce a stock solution of 100 mM and stored at −20°C. Capsazepine was dissolved in absolute ethanol to produce a stock solution of 100 mM. GSK1016790A (GSK101; [N -((1S)-1-{[4-((2S)-2-{[(2,4-dichlorophenyl)sulfonyl]amino-3-hydroxypropanoyl)-1-piperazinyl]carbonyl}-3-methylbutyl]-1-benzothiophene-2-carboxamide) was dissolved in dimethyl sulfoxide for a stock solution of 1 mM. Lidocaine hydrochloride, QX-314 chloride, QX-314 bromide, and N -methyl-d-glucamine chloride were directly dissolved in the extracellular solution (see Electrophysiology). The pH values of all solutions were adjusted to pH 7.4 with N-[2-hydroxyethyl]piperazine-N′-[2-ethanesulfonic acid] (HEPES) and NaOH. All control and test solutions were applied with an automated fast-switching multibarrel perfusion system (ValveBank8; AutoMate Scientific, Inc., Berkeley, CA).

All constructs were confirmed by DNA sequencing. Xenopus laevis  oocytes were injected with wild type human TRPV1 or TRPV4 messenger RNA (∼1 μg/μl) synthesized using the T7 mMessage mMachine kit (Ambion Canada, Streetsville, Ontario, Canada). After the injection, oocytes were incubated for 2–3 days for adequate expression. We transfected tsA201 cells with wild type plasmids of TRPV1 (10 μg) along with a reporter plasmid (CD8-pih3m; 2 μg) by the Ca²⁺-phosphate precipitation method. After being incubated for 7–8 h, the cells were replated in 35 mm glass-bottom culture dishes and used for experiments within 2–3 days. Transfection-positive cells were identified by immunobeads (anti-CD8 Dynabeads; Invitrogen Canada, Burlington, Ontario, Canada).

We recorded voltage-clamped TRPV1 and TRPV4 currents from oocytes with the two-electrode voltage-clamp technique using an oocyte clamp amplifier (model OC-725C; Warner Instrument Corp., Hamden, CT). Microelectrodes were pulled from borosilicate glass tubes (Frogy 1001; Harvard Apparatus, Holliston, MA). Ringer's solution (116 mM NaCl, 2 mM KCl, 1 mM MgCl², 0.5 mM CaCl², and 5 mM HEPES; adjusted to pH 7.4 with NaOH) was used as the standard external solution. In some experiments, NaCl was added to the standard Ringer's solution to produce an equiosmolar control solution compared to an applied drug-containing solution to rule out hyperosmolarity as a cause of TRPV1 activation. We perfused oocytes in a custom-made 30-μl bath chamber and used an automated fast-switching perfusion system (see Drugs and Chemicals) for drug application. The pCLAMP 10.0 software suite (Molecular Devices, Sunnyvale, CA) was used for data acquisition and offline analysis.

TRPV1 transient transfected tsA201 cells were loaded with the 1 μM acetoxymethyl ester form of fluo-4 (Molecular Probes, Eugene, OR) for 15 min and washed for an additional 10 min before recording. The dye was excited with an argon laser at 488 nm using an Olympus FV1000 confocal microscope (Olympus, Markham, Ontario, Canada). Analyses were performed offline using ImageJ software (Wayne S. Rasband, Special Volunteer, National Institutes of Health, Bethesda, MD).#Drugs were applied via  a micropipette (diameter, 100 μm) positioned 1 mm from the cells of interest via  a pressure ejection system. Data are expressed as fluorescence (F) or change in fluorescence (ΔF), divided by basal fluorescence (F⁰).

Calculations for statistical comparisons were performed with Prism 4 (GraphPad, San Diego, CA) and Sigma Plot 10 (Systat Software Inc., Chicago, IL). Continuous data were analyzed with repeated ANOVA (with Dunn's Multiple Comparison Test for post hoc  comparisons between individual groups) or paired Student's t  test as appropriate. All curves were fit using the Hill four-parameter, nonlinear regression model commonly used to characterize concentration dependence of pharmacologic responses. Statistical tests were two-tailed, and we considered a P  value ≤ 0.05 statistically significant. Data are given as mean ± SEM where appropriate.

To test the hypothesis that QX-314 directly activates TRPV1 channels, we examined Xenopus laevis  oocytes expressing human TRPV1 channels in a two-electrode voltage-clamp setup calibrated using the prototypical TRPV1 agonist capsaicin (fig. 1). Capsaicin produced no effects in oocytes injected with a saline control solution (data not shown). In oocytes injected with TRPV1 messenger RNA, administration of capsaicin produced a concentration-dependent reversible inward cationic current consistent with TRPV1 channel expression and activation (inset in fig. 1). All cells (n = 15) injected with TRPV1 mRNA responded to capsaicin. In our system, capsaicin demonstrated an activation range between ∼1 and 100 μM, with maximal activation occurring at ∼100 μM (fig. 1). The EC⁵⁰for capsaicin activation of TRPV1 was 19.2 ± 0.5 μM; consistent with previous findings,14,30the Hill coefficient was 2.1 ± 0.1 (Hill four-parameter, nonlinear regression, P < 0.001), indicating cooperativity between binding sites to open the TRPV1 ionic pore.31While activation occurred at higher concentrations than in some previous studies, it is possible that the observed differences in capsaicin sensitivity are attributable to differences in the expression system used (e.g ., oocytes vs . HEK293t cells), the TRPV1 isoforms studied (e.g ., rat vs . human), or the procedure of drug application (e.g ., we used an automated fast-switching perfusion system [see Materials and Methods and Drugs and Chemicals] and standardized drug applications to 10 s, ensuring a plateau would be reached, while limiting desensitization by extended capsaicin administration).32,33 

To investigate TRPV1 activation by QX-314, we applied increasing concentrations to Xenopus laevis  oocytes injected with TRPV1 messenger RNA. Before and after QX-314 administration, we applied capsaicin at a concentration producing a near-maximal response (50 μM) as a positive control and reference. Both QX-314 salts (bromide and chloride applied at 1–60 mM; see below) directly activated TRPV1 in a concentration-dependent manner. QX-314 chloride produced inward currents consistent with TRPV1 activation at 10–60 mM, with the largest effect occurring at 60 mM (fig. 2). Application of QX-314 bromide (≥30 mM) produced oocyte membrane blackening and cell death on contact with the solution (30 mM, n = 6/10 cells tested; not shown). Whereas the inward currents produced by QX-314 chloride generally were abolished after washout, at 60 mM some cells (n = 6/18) exhibited evidence of membrane disintegration upon administration, leading to cell death. For this reason, and to avoid possible nonspecific activation because of hyperosmolarity, we limited the range of QX-314 concentrations to 60 mM (see Discussion).15,34However, in experiments in which NaCl was added to the Ringer's control solution to render it equiosmolar with the solutions containing 60 mM QX-314 or lidocaine (see Materials and Methods, Electrophysiology), we observed no differences compared with standard Ringer's solution or effects attributable to hyperosmolarity in this range (data not shown).

Because QX-314 produced activation of TRPV1, we also tested a generic organic cation, N-  methyl-d-glucamine chloride, to determine whether the activation is unique to quaternary (i.e ., permanently charged) LAs such as QX-314 or if it represents a generic effect produced by high concentrations of permanently charged organic compounds. As shown in figure 2A, N-  methyl-d-glucamine chloride at a concentration equimolar to the highest QX-314 and lidocaine concentrations (60 mM) produced negligible TRPV1 activation. Higher concentrations of N  -  methyl-d-glucamine chloride (150 mM and 300 mM) evoked small inward currents (data not shown).

Consistent with the findings of a previous study,25lidocaine (1–60 mM) produced concentration-dependent TRPV1 activation. The experiment was performed in the same fashion as with QX-314, with applications of 50 μM capsaicin before and after lidocaine administration. Compared with QX-314 chloride, lidocaine produced larger inward currents at 10 and 30 mM but comparable effects at 60 mM, suggesting a higher efficacy (fig. 2, B and C). All cells perfused with lidocaine exhibited complete recovery, with no evidence of oocyte disintegration.

To test if the effects of QX-314 are exclusive to TRPV1 or also apply to other transient receptor potential isoforms, we studied its actions on TRPV4, the closest TRPV1 homolog (see Discussion). Using the potent TRPV4 agonist, GSK101 (500 nM), as control, we found that QX-314 chloride produced no activation of TRPV4 channels (60 mM; n = 7; fig. 3).

To investigate if QX-314 activates TRPV1 channels through mechanisms similar to that of capsaicin, we examined the effects of the well-known and well-characterized competitive TRPV1 antagonist, capsazepine, on QX-314- and lidocaine-evoked currents (fig. 4). The experiments were carried out in a manner similar to those in figure 2, with 50 μM capsaicin applied at the beginning and end of each experiment. As shown in figure 4, A and B, QX-314–evoked currents were effectively blocked by capsazepine. Similarly, capsazepine produced near complete blockade of TRPV1 activation by lidocaine (100 μM; fig. 4, C and D).

To rule out the possibility that the observed TRPV1 channel activation by QX-314 was specific to the Xenopus laevis  oocyte expression system, we sought to verify our results on TRPV1 channels transiently transfected in the mammalian tsA201 cell line. We determined expressed TRPV1 activity by measuring fluo-4 acetoxymethyl ester Ca²⁺transients with laser-scanning confocal microscopy. We used 10 μM capsaicin as control and found that fluorescence from Ca²⁺entry into the cytosol was readily apparent in tsA201 cells (data not shown). Application of 60 mM QX-314 chloride produced a robust increase in fluorescence compared with basal (background) fluorescence levels (fig. 4E). The increase in fluorescence was similar for 10 μM capsaicin and 60 mM QX-314 (data not shown). The transient increase in cytoplasmic Ca²⁺evoked by QX-314 (or capsaicin) was nearly completely inhibited by capsazepine (fig. 4E).

A previous study suggested that QX-314, like lidocaine, can inhibit capsaicin-evoked TRPV1 currents.28However, the concentrations used in these experiments were comparatively high (30 mM), possibly underscoring the inhibitory potency of QX-314 because our results clearly show direct activation of TRPV1 by QX-314 in the millimolar range. Consequently, we proceeded to determine whether QX-314 might also inhibit TRPV1 currents at subactivating concentrations. To test this possibility, we investigated the effects of micromolar QX-314 concentrations on currents activated by capsaicin at a concentration near its EC⁵⁰(17 μM). As illustrated in figure 5, the data demonstrate that QX-314 blocked inward currents in a concentration-dependent and reversible manner, with 79 ± 8% of current blocked at 30 μM and 93 ± 4% blocked at 300 μM (n = 15). The calculated IC⁵⁰was 8.0 ± 0.6 μM, indicating that QX-314 is a potent inhibitor of TRPV1 channels at subactivating concentrations.

Our results show that the quaternary lidocaine derivative, QX-314, exerts biphasic effects on TRPV1 channels, inhibiting capsaicin-evoked TRPV1 currents at lower (micromolar) concentrations and activating TRPV1 channels at higher (millimolar) concentrations. From our data, it is apparent that QX-314, administered in the millimolar range, activates TRPV1 channels in a concentration-dependent manner. Although we were unable to reach saturation with QX-314 in our experiments and construct a full concentration-response curve on TRPV1 activation, our results indicate that QX-314 is a slightly less potent (and/or less efficacious) TRPV1 activator than is lidocaine, which previously was found in human embryonic kidney (HEK293t) cells to produce TRPV1 activation with an EC⁵⁰of 12 mM.28Of note, the authors found evidence that 100 mM lidocaine (a maximally effective concentration) induced seal breaks in the cells. Somewhat similarly, we observed in our study that QX-314 concentrations ≥60 mM (corresponding to approximately a 2% solution) produced oocyte disintegration and cell death, raising the possibility of a cytotoxic effect of these high concentrations.

The results of this study with QX-314 may provide a basis for modification of the proposed molecular model for LA activation of TRPV1, which previously was thought to occur through a hydrophobic pathway to activate a domain in the cytosol.28It has been suggested that LAs interact with a domain that is similar, but not identical, to the vanilloid-binding region composed of transmembrane domains 3 and 4 and the respective cytosolic interfaces.28,30,35Thus, it is noteworthy that extracellular application of the permanently charged molecule, QX-314, produced effects similar to that of lidocaine on TRPV1. A possible explanation for this observation is that the other regions of the channels may be involved in LA binding and that binding domains for vanilloids and lipophilic LAs are distinct.15,28Previous investigations with HEK293t cells found that TRPV1 was not activated by 30 mM QX-314 (corresponding to approximately a 1% solution) compared with robust lidocaine activation at an equimolar concentration.28However, we have demonstrated that 60 mM QX-314 activates TRPV1 channels expressed in Xenopus laevis  oocytes and tsA201 cells. Importantly, in both cell types, activation was inhibited by capsazepine, indicating that the two LAs use a similar activation pathway, one that is also similar to that of the prototypical agonist, capsaicin. Thus, it would appear that QX-314 produces activation through a molecular mechanism similar to that used by vanilloids such as capsaicin.36,37 

If QX-314 evokes TRPV1-mediated currents in oocytes and tsA201 cells, why did a previous study not observe QX-314-dependent activation in HEK293t cells (although QX-314 inhibited TPRV1-evoked currents)? First, the results came from experiments with a single concentration of both lidocaine and QX-314, precluding a thorough concentration-response analysis. Thus, a biphasic action of QX-314 may have been missed. Second, the HEK293t cells in the previous study were transfected with rat TRPV1 channels,28whereas in the current study, we expressed human TRPV1 channels in oocytes and tsA201 cells, which may explain some of the observed differences. Finally, TRPV1 channels may adopt a different molecular conformation in oocytes and tsA201 cells than in HEK293 cells, allowing the LA-binding domain to be exposed to a highly hydrophilic compound such a QX-314. However, this is unlikely because we also recorded a significant increase in cytoplasmic Ca²⁺concentrations in TRPV1-transfected tsA201 cells exposed to 60 mM QX-314 (see fig. 4).

Mammalian tissues differentially express a variety of transient receptor potential channel isoforms.38,39We sought to investigate if QX-314 would evoke currents in transient receptor potential isoforms closely related to TRPV1, the closest homolog being TRPV4. The TRPV4 channel has ∼40% homology with TRPV1, particularly in the area of the ion pore and selectivity filter.40,41Like TRPV1, it is activated by a wide range of physical and chemical stimuli, including synthetic phorbol derivative 4α-phorbol-12,13-didecanoate,42arachidonic acid metabolites,43moderate warmth (more than 27 °C),42hypotonic cell swelling,44pressure,45membrane stretch,46and sheer stress.47With the use of the novel TRPV4 agonist, GSK101, as a positive control,48QX-314 produced no activation in these channels. This observation suggests that QX-314 and lidocaine may be activating TRPV1 channels through a TRPV1-specific mechanism, possibly involving a region similar to or near the vanilloid-binding domains. However, we cannot exclude the possibility that the conformation TRPV4 presents in Xenopus laevis  oocytes does not allow QX-314 access to bind and activate the channel.

In terms of potential clinical significance for therapeutics and toxicity, the findings of the current study might help explain our recent in vivo  observations of severe irritation associated with intrathecal QX-314 administration in mice.29In general, pain upon injection of LAs is a well-known phenomenon and often a source of discomfort for patients undergoing local anesthesia.49One possible mechanism may be the low pH of some LA formulations,50because TRPV1 is considered the main transducing receptor system for proton-induced excitation of nociceptive sensory neurons.28,51,52However, the drug-containing solutions in our study were titrated to a pH of 7.4, and previous studies have suggested that LAs themselves may produce pain upon injection by directly activating TRPV1.28While we found that both QX-314 and lidocaine activate TRPV1 at clinically relevant concentrations, lidocaine-evoked pain rapidly ceases as nerve conduction block sets in, a feature that is unlikely shared by QX-314 because of its significantly longer LA onset time.7This may explain why intrathecal lidocaine does not produce the sustained severe acute irritation associated with QX-314.29That said, lidocaine is well known to be neurotoxic, and on the basis of the current study, it would seem possible that Ca²⁺influx mediated by TRPV1 activation53–55contributes to LA-induced cell death by inducing necrotic and apoptotic mechanisms.56,57The cellular degeneration caused by QX-314 chloride (60 mM) and bromide (≥30 mM) in this study may well relate to the adverse intrathecal effects observed in vivo . In this regard, QX-314 appears to share features of the toxicity profile of other quaternary ammonium compounds (e.g ., tetraethylammonium-C12,58,N-  phenylethyl amitriptyline, N -propyl amitriptyline, N-  propyl doxepin,59–61and tonicaine [N -phenylethyl lidocaine]62–64), all of which act as LAs but produce severe neurotoxicity in vivo . Collectively, the available data raise the possibility that the quaternary ammonium function and associated permanent positive charge confer specific structure-activity properties that give rise to an adverse effect profile distinct from that of amphipathic LAs. In this regard, our current and previous29laboratory research findings raise questions about the safety and feasibility of QX-314 use in humans, particularly in central neuraxial blockade. Clearly, the answers require more research.

The action of QX-314 on TRPV1 channels varied greatly with the administered concentration. In contrast to the effects at millimolar concentrations, QX-314 at lower, subactivating (micromolar) concentrations potently inhibited capsaicin activation of TRPV1 in vitro , with an IC⁵⁰of 8 μM and near complete blockade reached at 300 μM. Lidocaine previously has been shown to block rat TRPV1 channels expressed in HEK293t cells, albeit with a substantially lower potency (IC⁵⁰, 45 mM; corresponding to approximately a 1.5% solution).28In addition, TRPV1 inhibition from the intracellular pore has been reported with other quaternary ammoniums, including tetraethylammonium, tetrapropylammonium, tetrabutylammonium, and tetrapentylammonium.65While the specific molecular mechanisms for the complex effects of extracellularly applied QX-314 on TRPV1 channels remain to be elucidated, our observations provide novel insight into the interactions between this quaternary lidocaine derivative and TRPV1 channels. Because TRPV1 channels play a central role in peripheral nociceptive transduction, they represent attractive potential therapeutic targets for the management of acute pain and chronic pain syndromes. Taken together, our current findings and the results of recent studies indicate that the TRPV1 channel represents both a novel LA target and pathway for these agents to reach their intracellular binding site on the voltage-gated Na⁺channel.8,9,28 

In conclusion, we show in this work involving experiments on Xenopus laevis  oocytes and tsA201 cells that the quaternary lidocaine derivative, QX-314, behaves as a biphasic regulator of expressed human TRPV1 channels. At low micromolar concentrations, QX-314 inhibits TRPV1 activation in the presence of capsaicin, whereas at higher concentrations in the millimolar range more relevant to clinical regional anesthesia and neural blockade, QX-314 is a TRPV1 channel agonist similar to lidocaine.28Our results demonstrating dichotomous actions on a specific molecular target not only describe what we believe to be a novel property of quaternary LAs, but also may provide a new molecular explanation for LA-induced adverse effects in regional anesthesia.

The authors thank Sebastian Brauchi, Ph.D. (Assistant Professor, Universidad Austral de Chile, Campus Isla Teja, Valdivia, Chile), for generously providing the transient receptor potential channel clones; Guy Tanentzapf, Ph.D. (Assistant Professor, Department of Cellular & Physiological Sciences, The University of British Columbia, Vancouver, British Columbia, Canada), for his generous assistance with the Ca²⁺imaging experiments; Ana Niciforovic, Ph.D. (Technician, Department of Anesthesiology, Pharmacology & Therapeutics, The University of British Columbia), for technical help; and Bernard A. MacLeod, M.D., F.R.C.P.C. (Associate Professor, Department of Anesthesiology, Pharmacology & Therapeutics, The University of British Columbia), for his valuable comments and inspiration.