|Home | About | Journals | Submit | Contact Us | Français|
Pain results from the complex processing of neural signals at different levels of the central nervous system, with each signal potentially offering multiple opportunities for pharmacological intervention. A logical strategy for developing novel analgesics is to target the beginning of the pain pathway, and aim potential treatments directly at the nociceptors — the high-threshold primary sensory neurons that detect noxious stimuli. The largest group of receptors that function as noxious stimuli detectors in nociceptors is the transient receptor potential (TRP) channel family. This Review highlights evidence supporting particular TRP channels as targets for analgesics, indicates the likely efficacy profiles of TRP-channel-acting drugs, and discusses the development pathways needed to test candidates as analgesics in humans.
Primary sensory neurons are the interface of the nervous system with the external and internal environments that exist outside and inside our bodies. These neurons have a cell body located in the dorsal root ganglion (DRG), a peripheral axon that innervates tissues such as skin and whose terminals react to sensory stimuli, and a central axon that enters the spinal cord to transfer information to the central nervous system (CNS) by synaptic communication. A major function of the sensory apparatus is to detect potentially damaging stimuli and thereby warn of the risk of injury. This key survival tactic has evolved so that the sensations evoked by noxious stimuli are intensely unpleasant and consequently can be avoided the next time they are encountered. However, the threshold for eliciting pain must be high enough so that most activities can be carried out largely pain-free, but sensitive enough so that an alert can be given immediately if an injury is impending. The first neural mediator of this crucial alarm system is the nociceptor1.
The detection of noxious stimuli by nociceptors elicits nociceptive pain, such as that elicited after touching a hot object or in response to an intense pinch (FIG. 1). However, in various pathological conditions pain can occur in the absence of a noxious stimulus, in response to normally innocuous stimuli (allodynia), and with an exaggerated response to a noxious stimulus (hyperalgesia). After tissue injury the inflammatory response sensitizes nociceptors so that their threshold for activation drops and their responsiveness increases, and this contributes to pain hypersensitivity at an inflamed site, a distinct component of inflammatory pain2 (FIG. 1). Following peripheral nerve damage, nociceptors may begin to fire ectopically, and in this way contribute to the spontaneous component of neuropathic pain3 (FIG. 1).
Pharmacological intervention to reduce pain can produce analgesia by either decreasing excitation or increasing inhibition in the nervous system. Opioids, for example, decrease neurotransmitter release presynaptically and hyperpolarize neurons postsynatically in the spinal cord, brainstem and cortex4. Sodium-channel blocking and potassium-channel opening anticonvulsants reduce excitation throughout the nervous system, whereas amine uptake inhibitors potentiate the actions of inhibitory transmitters in the spinal cord and brain5,6. A problem associated with centrally acting drugs — such as opioids, antidepressants, anticonvulsants and sodium-channel blockers that target receptors/channels that are widely expressed — is a higher risk of adverse effects such as sedation, dizziness, somnolence or loss of cognitive function. An alternative strategy for developing novel analgesics is to target the very beginning of the pain pathway and aim treatment directly at receptors and ion channels that transduce noxious stimuli at the peripheral terminals of nociceptors into electrical activity. These nociceptive ion channels, which essentially define the characteristic functional properties of nociceptors, are selectively or mainly expressed in these neurons7–9, thereby potentially reducing the side-effect profiles of drugs that specifically act on them. The largest group of noxious stimulus detectors is the transient receptor potential (TRP) channel family7–9, which this Review will highlight, with a particular emphasis on TRPV1 and TRPA1 as targets for analgesics (FIG. 2).
The first suggestion that TRP channels are key receptors involved in sensory transduction emerged from the identification of the original TRP ion channel from Drosophila melanogaster10. The name of the ion channel is derived from a fly mutant that displays a transient instead of a sustained response to bright light. TRP ion channels from various species are important for vision, olfaction, taste, mechanosensation, osmoregulation and thermosensation7. A role of TRP channels specifically in pain and thermosensation was first suggested by the finding that mammalian TRPV1 is activated by both noxious heat and capsaicin, the active ingredient of chilli peppers11. Importantly, the cloning and characterization of several close and distant relatives of TRPV1, which are activated at distinct thresholds of cool and warm temperatures, established a general role for TRP channels in thermosensation and nociception: TRPV2 (REF. 12), TRPV3 (REFS 13–15), TRPV4 (REFS 16,17), TRPM8 (REFS 18,19) and TRPA1 (REF. 20).
Although there is little amino-acid conservation among distant TRP channels, they share a similar architecture of six-transmembrane domains with cytoplasmic amino and carboxy termini. TRP channels are thought to function as tetramers, mostly as homomers21. Six of the 28 TRP channels from the three distinct TRP family subtypes are activated by temperature (TRPV1–4, TRPM8 and TRPA1), and are expressed in sensory neurons or in skin keratinocyte cells, which are peripheral targets of these nerves8. Three other TRP channels (TRPM2, TRPM4 and TRPM5) are strongly modulated by warm temperatures; however, the lack of expression in nociceptor neurons argues against a role in nociception22,23.
One way to elucidate the function of somatosensory receptors is to examine the behavioural consequences of specifically activating them by chemical ligands or thermal stimuli — a gain-of-function approach.
TRPV1 was identified as a capsaicin- and heat-activated ion channel, and analysis of TRPV1-deficient mice established that capsaicin acts entirely through TRPV1 (REFS 24,25). The sensory qualities of capsaicin are described as burning, itching, piercing, pricking and stinging; sensory qualities that are clearly related to pain26,27. Indeed, most mammals (excluding those humans who find TRP activation enhances culinary experiences) avoid capsaicin and other pungent compounds. In addition to capsaicin, TRPV1 is activated by spider toxins, noxious heat and is modulated by low pH (a common consequence of inflammation)11,28. Therefore, TRPV1 can be defined as a polymodal nocitransducer.
Many pungent chemicals, excluding capsaicin, activate TRPA1. The list includes several compounds found in food, such as isothiocyanates (horseradish, mustard), cinnamaldehyde (cinnamon) and allicin (garlic)29–32. Such pungent compounds are all electrophiles that activate TRPA1 through covalent modification of reactive amino acids such as cysteines33,34. Therefore, this form of chemical TRPA1 modulation is not dictated by the structure of the activating molecule per se, but rather by its chemical reactivity. Indeed, all potent cysteine-reactive chemicals tested seem to activate TRPA1, irrespective of overall shape33,34. Other TRPA1 activators include the alkylating agent iodoacetamide; the environmental irritant acrolein, found in cigarette smoke and an industrial pollutant; formaldehyde, the most commonly used agent to assay chemical nociception in rodents; acetaldehyde, an intermediate substrate of ethanol metabolism; and the endogenous 4-hydroxynonenal (4-HNE)35–38. Reactive oxygen and nitrogen species have also been described to activate TRPA1 (REFS 39–41). 4-HNE is of particular interest as an activator of TRPA1, as it is produced by lipid peroxidation in cells and may be responsible for the pathological effects of oxidative stress. Similar to other TRPA1 activators, 4-HNE can bind to cysteine, histidine and lysine residues in proteins via a Michael addition reaction42 with an EC50 value of 13 μm. This is physiologically relevant as oxidative stress causes accumulation of 4-HNE at concentrations of 10 μm to 5 mm in membranes42. Although many of these chemicals irreversibly react with cysteines, it has been argued that the covalent modification of cysteines in TRPA1 by mustard oil is rapidly reversible34,43. However, the half-life of isothiocyanate–cysteine adducts is in the order of 1 hour at physiological pH and temperature44. Furthermore, the successful labelling of mustard oil alkyne adducts on TRPA1 by click reaction suggests that these adducts remain stable, as the click reaction itself takes 1 hour to complete33. TRPV1 can also be activated by cysteine modification by allicin, although it is less promiscuous than TRPA1, and many cysteine modifiers do not activate this channel45.
There is strong evidence that the covalent binding mechanism of TRPA1 activation is physiologically relevant. Although these reactive chemicals modify proteins non-discriminately, the nociception they cause is dramatically reduced or eliminated in Trpa1−/− mice36–38. Therefore, TRPA1 is a key sensor of chemical damage in vivo and could be an exciting target for treating various pain states as well as bronchial hyperresponsiveness to inhaled irritants in asthma. In addition to reactive chemicals, TRPA1 is activated in some experimental settings by noxious cold temperatures20,31,46,47 (although this remains controversial) and by calcium46,48, and is, therefore, also a polymodal nocisensor. Calcium is not a noxious stimulus per se, nevertheless, the calcium activation of TRPA1 in nociceptive neurons raises the possibility that TRPA1 acts as an amplifier of other signals that increase intracellular calcium, including TRPV1 (FIG. 3). However, the role of calcium on TRPA1 modulation is complicated by the fact that it also causes a desensitization of this channel49.
TRPM8 is activated by innocuous cold temperatures and by cooling compounds such as menthol18,19. These TRPM8-activating stimuli are not, however, noxious, although they could indirectly influence nociceptive signalling and may contribute to cold allodynia in inflammatory and neuropathic pain (see knockout studies below). Specific activators of the other thermoTRP channels (TRPV2, TRPV3 and TRPV4) with a clear nociceptive quality have not yet been identified (but see below for some tentative evidence)8.
TRPV1 and TRPA1 are co-localized in a subgroup of DRG and trigeminal neurons that constitute polymodal nociceptors, which are specialized sensors of a variety of noxious stimuli20,31. However, unambiguous expression of the other thermoTRP channels in nociceptors has not been established. TRPM8, for example, is mainly expressed in a separate group of sensory neurons that respond to innocuous cool temperatures (although a minor overlap between TRPV1-expressing neurons is observed)50,51. TRPV2 marks a very poorly characterized subpopulation12. Low levels of TRPV3 and TRPV4 expression have been reported in DRG neurons (including in humans), but seem to be more prominently expressed in keratinocytes13,14,16,52–54. Among the thermoTRP channels, TRPM8 and TRPA1 seem to be exquisitely specific to sensory neurons, whereas the other TRP channels are expressed in several neuronal and non-neuronal tissues8.
Although TRP channels are the archetypal transducer receptor — converting thermal and chemical stimuli into electrical activity on the peripheral terminals of sensory neurons — they are also located on axons, which respond to noxious heat stimuli55. TRP channels are also found on the central terminals of nociceptors that make synaptic contact with neurons in the dorsal horn of the spinal cord56. Activation of central terminal TRPV1 by capsaicin, and of TRPA1 by mustard oil applied to isolated spinal cord preparations, results in an increase in synaptic release of both glutamate and neuropeptides57–59. These results indicate a possible role for the TRP channels as synaptic modulators, both directly (by controlling calcium influx and thereby synaptic vesicle release) and indirectly (by producing a depolarization of the terminal). However, what activates central TRP channels — which could include temperature, a chemical ligand or intracellular calcium — is unknown. One set of candidates for TRPV1 are the endocannabinoids, including anandamide, which has a modulatory action on the channel60. For TRPA1 it could conceivably be an endogenous oxidation product, such as 4-HNE, or the coupling of bradykinin through its B2 G-protein-coupled receptor (GPCR) with TRPA1 (REF. 61). Bradykinin is produced in the spinal cord in response to noxious inputs and causes an increase in glutamate release62. A recent surprising finding is that in the hippocampus, TRPV1 receptor activation selectively modifies synapses into inhibitory interneurons to produce long-term depression. This TRPV1-mediated long-term depression may have a role in long-term changes in physiological and pathological circuit behaviour during learning and epileptic activity, and might reveal unexpected side effects or uses for TPV1 antagonists63. TRPV1 also seems to be tonically activated by endovanilloids in brainstem neurons where it may modulate descending antinociceptive pathways64. The specificity of putative TRPV1 actions beyond primary sensory neurons, where its expression levels are by far the highest7–9, need to be confirmed.
Inflammatory signals and nerve injury alter TRPV1 expression and function by multiple mechanisms — which include transcriptional and translational regulation, post-translational changes and altered trafficking — and in this way contribute to pathological pain states by increasing sensitivity to noxious stimuli2 (peripheral sensitization) (FIG. 3). Retrograde transport of nerve growth factor (NGF) acts via p38 to increase translation of TRPV1 in the cell body during inflammation, resulting in an increased transport of TRPV1 to peripheral terminals65. NGF also has more immediate actions that increase TRPV1-mediated currents in nociceptor peripheral terminals. NGF acts on tryosine receptor kinase A (TRKA; also known as NTRK1) to activate a phosphoinositide 3-kinase (PI3K)–SRC kinase signalling pathway to phosphorylate intracellular stores of TRPV1, which leads to the insertion of TRPV1 into the cell membrane66. NGF also alters the threshold of TRPV1 via PI3K and p42/p44 mitogen-activated protein kinases67. Many inflammatory mediators that sensitize TRPV1 do so via protein kinase Cε (PKCε)68, which enables TRPV1 to be a major integrator of diverse inflammatory signals on nociceptors (FIG. 3). Indeed, it is this body of evidence that points to the possible efficacy of TRPV1 antagonists in diverse inflammatory pain conditions, including post-surgical pain, cystitis, ulcerative colitis, oesophagitis, asthma and rheumatoid arthritis. Phospholipase C (PLC) and the membrane phospholipid phosphatidylinositol 4,5-bisphosphate it produces, seem to have both inhibitory and activating effects on TRPV1 (REF. 69).
In contrast to inflammation and bone cancer, in which TRPV1 levels increase substantially70, expression of TRPV1 in neurons decreases in response to peripheral axonal damage71,72. TRPV1 levels are reduced in the skin of patients with diabetic neuropathy but are increased in intraepidermal fibres in some patients with pain hypersensitivity and proximal to some forms of traumatic nerve injury, where they may accumulate because transport to the periphery is disrupted71. Unmyelinated axons respond to noxious heat with a threshold similar to their peripheral terminals and to heterologous TRPV1 at ~41 °C55. It is not known whether TRPV1 located in axonal membranes and in neuroma are activated by endovanilloids or become sensitized so that their threshold falls to body temperature in neuropathic pain. The contribution of TRPV1 to neuropathic pain may include sensitization of intact peripheral terminals, ectopic activity in injured axons or a role in modulating transmitter release. Interestingly, TRPV1 antagonists reduce pain sensitivity in some models of neuropathic pain73, although a knockout of TRPV1 generally has no effect24,25. However, global knockout of a gene may not always be a useful predictor by itself of a role zin disease states (BOX 1).
How useful are transgenic models in predicting therapeutic benefit? Approximately 50% of the molecular targets of the top 100 selling drugs have been knocked out in mouse transgenic models, as have many targets of drugs in full development. In both cases approximately 85% of knockout phenotypes showed a sound biological rationale for the disease indication, with a direct correlation between knockout phenotype and the effect of drug. This therefore demonstrates that knockout models are likely, in most cases, to provide a productive source of validated targets for future drug development142.
Nevertheless, with respect to pain phenotyping, mouse gene knockouts are subject to many potential pitfalls, including functional compensation by related channels and receptors, and significant differences between mouse strains143. Significant advances have been made to improve knockout mouse technology that may help improve their utility for target validation. Transgenic mouse models need to be generated on a specific genetic background for phenotypic analysis, primarily strain C57BL/6. Until quite recently, however, routine model generation has only been possible through gene targeting in the robust 129 embryonic stem cells, followed by successive backcrossing onto the C57BL/6 background. This adds up to 12 months to model generation timelines and is resource intensive.
It is now possible to generate transient receptor potential (TRP) knockout mice by gene targeting in C57BL/6 embryonic stem cells, which obviates the need for a year of backcrossing and allows behavioural phenotyping on a pure genetic background144. Derivation of transgenic mice using C57BL/6 cells is becoming routine practice in many laboratories145. This advance coupled with advances in conditional and tissue specific models will improve the timeliness, utility and application of future TRP transgenic models.
The extent to which inflammatory signals also upregulate and sensitize other thermoTRP channels has been less extensively investigated. TRPA1 is activated by high concentrations and sensitized by low concentrations of bradykinin downstream of G-protein-coupled signalling32,61. The bradykinin sensitization is modulated by both PLC and PKA61. TRPA1 is also activated by protease-activated receptor 2 (PAR2; also known as F2RL1) through PLC74. There is also evidence that TRPA1 is transcriptionally regulated. Nerve injury, complete Freund’s adjuvant and NGF increase TRPA1 expression and activity in sensory neurons75,76 (see also loss-of-function studies below). Studies have similarly highlighted the sensitizing effects of inflammatory signals such as prostaglandin E2 (PGE2) and PAR2 on TRPV4 activity77,78. TRPV3 is sensitized by repeated stimulation, and calcium plays an important role in this process13,79. Furthermore, TRPV3 activity is potentiated by PLC-dependent GPCR activation80. However, a role for TRPV3 in the sensitization of nociceptors has not yet been shown in vivo. In contrast to the sensitizing effects of inflammatory signals on other thermoTRP channels, these signals mainly suppress TRPM8 function. For example, PKC activation and phosphatidylinositol 4,5-bisphosphate depletion (which occur in response to inflammatory signals such as bradykinin) abolish or reduce TRPM8 activity81–83. TRPM8 responses are abolished by inhibitors of the calcium-independent form of phospholipase A2, whereas lysophospholipids — one of the enzyme’s products — activate the ion channel84,85. Whether this pathway is involved in nociceptor sensitization in vivo is not clear.
Trpv1−/− mice demonstrate a profound requirement for TRPV1 for thermal hyperalgesia in response to inflammatory signals, although a reduction but not elimination of acute thermal pain is also observed in these mice24,25 (BOX 2). Since these initial findings, analysis of Trpv1−/− mice has also established a role of TRPV1 in bladder function and inflammation, serum osmolarity, diabetes, pancreatitis, cough, arthritis, anxiety, and fever86–91. Acute pharmacological block of protein activity in adult animals sometimes deviates from the effects of germline ablation of the encoding gene. For example, pharmacological inhibition of TRPV1 induces hyperthermia in wild-type mice in a TRPV1-dependent manner, whereas Trpv1−/− mice do not display this phenotype92,93. Similarly, many TRPV1 antagonists seem to reduce mechanical nociception, whereas Trpv1−/− mice have normal mechanical responses (reviewed in REF. 94). Clearly, compensatory mechanisms are at play in mice that lack TRPV1 throughout development. Overall, the genetic and pharmacological studies on TRPV1 establish potential clinical uses for targeting TRPV1 for pain indications, but also raise concerns regarding potential side effects, such as a diminished response to damaging heat stimuli, altered body temperature and a reduction in the perception of taste.
Below is a list of the published phenotypes of Trpv1−/− knockout mice:
What is the anatomical site of action of TRPV1 antagonists — the peripheral or central end of the nociceptor? Intrathecal administration of capsazepine or iodoresiniferatoxin reduces intraplantar formalin-evoked pain-related behaviour95. By contrast, when a TRPV1 antagonist, SB-366791, is applied to an isolated spinalcord preparation there is a reduction in glutamatergic transmission after peripheral inflammation, implying that in these situations the central TRPV1 receptor is tonically active96. Furthermore, a TRPV1 antagonist with poor CNS penetrance (A-795614) has equal potency to one with good CNS penetration (A-784168) for peripheral capsaicin-evoked pain-like behaviour and thermal hyperalgesia after peripheral inflammation, implying action at the peripheral terminal97. However, the compound with greater CNS penetration is more potent when given systemically for mechanical allodynia and osteoarthritic pain than the compound that does not cross the blood–brain barrier. These findings suggest, perhaps somewhat surprisingly, that targeting TRP channels on the central terminals of nociceptors is required to achieve maximal analgesic benefit. The challenge for the pharmaceutical industry, therefore, is to identify the next generation of compounds that have the ability to cross the blood–brain barrier, while retaining the necessary potency, selectivity and developability properties and which minimize adverse effects. Although these would be centrally acting compounds they would still act primarily on nocicptors, albeit at their central and not peripheral terminal, and in consequence may retain a high therapeutic index.
Two groups have independently shown that Trpa1−/− mice have strong deficits in somatosensory chemosensation98,99, thereby confirming the role of this receptor in sensing noxious stimuli. One group observed deficits in acute nociceptive mechanosensation and cold thermosensation in Trpa1−/− mice98. The mechanosensory deficit was not examined by the other group; however, an independent study was able to reproduce the specific mechanosensory deficit100. For cold thermosensation this result is disputed and more investigations including analysis of Trpm8/Trpa1 double mutant mice are required, as the presence of TRPM8 might be masking any subtle phenotype present in TRPA1-deficient mice. However, in vivo antisense experiments have shown a requirement for TRPA1 for cold hypersensitivity in response to inflammation and injury in rats76. A selective small-molecule inhibitor of TRPA1 that is ineffective in Trpa1−/− mice reverses complete Freund’s adjuvant-induced mechanical hyperalgesia in wild-type mice37,100. These experiments collectively suggest that TRPA1 plays a part in establishing or maintaining mechanical hyperalgesia, and that compensation in Trpa1−/− mice may mask this requirement. In addition, Trpa1−/− mice have deficiencies in oxidant-induced respiratory depression, nasal obstruction, sneezing, cough and pain, emphasizing that TRPA1 activation has consequences beyond pain perception101 (BOX 3). More studies are needed to elucidate the potential effectiveness of targeting TRPA1 in various pain indications as well as in asthma and other respiratory problems.
Below is a list of the published phenotypes of transient receptor potential (TRP) channel knockout mice that have been generated. Further information on the effect of TRP channel knockouts on pain phenotype can be obtained from the Pain Genes database (see Further information).
Overall, these early studies on TRPV1 and TRPA1 underscore the importance of considering both acute pharmacological blockade and genetic ablations in assigning gene function and potential clinical utility.
Analysis of Trm8−/− mice clearly establishes that this ion channel is the major sensor of peripheral innocuous cool temperatures50,102,103 (BOX 3). However, sensing noxious cold nociception is only partially affected in these mice, which suggests the existence of at least one other receptor. A role of TRPM8 in cold allodynia after nerve injury is suggested in rats and in mice102,104. Paradoxically, TRPM8 activation has also been shown to have analgesic properties50,105. It is proposed that TRPM8 expression in functionally distinct DRG neurons can account for multiple (even opposing) roles in somatosensation106. TRPM8 is mainly expressed in DRG neurons that respond to cool temperatures, but most do not express any nociceptive markers. About 10–20% of TRPM8-positive neurons also express TRPV1 and are most likely to be nociceptors51,107. Therefore, TRPM8-expressing neurons could send nociceptive as well as cool signals to the spinal cord, depending on the context. Whether TRPM8 is a useful target for treating cold allodynia in patients with neuropathic pain or for other indications is difficult to assess at this time. Pharmacological disruption of TRPM8 in adult animals should shed light on this.
Trpv4−/− mice display various defects including impaired thermal, mechanical and osmotic responses, which are mostly consistent with the different known modes of TRPV4 activation77,108–115 (BOX 3), but are also incontinent and deaf. These studies also validate that TRPV4 is involved in nociceptor sensitization. One outstanding question is whether keratinocytes or sensory neurons are the site of TRPV4 action53,54,78,112. Interestingly, TRPV4 may have a particular role in colonic afferents where it is involved in mechanotransduction, and is expressed in nerve fibres of patients with inflammatory bowel disease116. Given the wide-ranging phenotypes described for TRPV4-deficient mice, it is unclear whether systemic TRPV4 blockade for analgesia would be accompanied by severe side effects110, and whether a non-absorbable TRPV4 blocker could be used for colonic pain. Once again, pharmacological blockade in adult animals should directly address these issues. TRPV4 is also involved in the generation of pulmonary oedema in response to increased microvascular pressure and this can be blocked by cyclic GMP117. Therefore, inhaled TRPV4 antagonists, like those for TRPV1 and TRPA1, may have utility in respiratory diseases.
Acute noxious heat thermosensation is diminished in Trpv3−/− mice, but there is no evidence to date for a role of TRPV3 in inflammatory and neuropathic pain118. Trpv2−/− mice have not yet been described. Testing knockout mice for these TRP channels in clinically relevant pain models, and in particular identifying and testing specific antagonists against all thermoTRP channels for efficacy in pain states, is necessary to validate which of these ion channels are potential targets for novel analgesics (BOXES 2,3).
Will blocking TRPV1 and TRPA1 or other thermoTRP channels be useful for pain indications? One could argue that such alarm signals for tissue damage have evolved for protective effects, and blocking them would therefore not be desirable. However, in pathological states, changes in the chemical milieu of nerve endings could overactivate and sensitize the pain pathway through nociceptive TRP channels (FIG. 3), thereby leading to pain hypersensitivity in many disease states119. For example, patients suffering from diabetic neuropathy have elevated levels of reactive chemicals that are known TRPA1 activators120,121. Could this explain the ectopic pain present in this condition and in other painful neuropathies? Does the threshold of TRPV1 expressed in unmyelinated axons ever drop to a point that it is activated at body temperature55? These questions will hopefully be addressed within the next 2–5 years, both in animal models and in Phase IIA proof-of- concept trials with suitable specific antagonists. These specific antagonists will also help address whether TRP antagonists need to be targeted only to the periphery to produce maximum analgesia (FIG. 4).
ThermoTRP channels are more than just acute damage sensors that mediate acute nociceptive pain; they are also important players in inflammatory pain and possibly also neuropathic pain. Indeed, TRPV1 and TRPA1 activation does not only send an afferent (sensory) signal to the CNS but also an efferent (motor) signal via secretion of inflammatory agents such as substance P and calcitonin gene-related peptide (CGRP)9 (FIG. 2). This efferent signal at the peripheral terminal causes local neurogenic inflammation with vasodilation and oedema owing to increased capillary permeability9. TRPV1-expressing afferents may trigger the auto immune inflammatory response in type 1 diabetes91. As nociceptive TRP channels are themselves targets of inflammatory signals (FIG. 3), they are also mediators of peripheral sensitization. The challenge clinically will be to minimize disruption of the acute damage sensor role while maximizing reduction of the TRP channel role in pathological conditions. Given their participation in pain transduction and inflammatory signalling, their sensitization properties and their expression pattern, targeting these ion channels could possibly ameliorate many inflammatory conditions including arthritis, some forms of neuropathic pain such as diabetic neuropathy, itch, asthma, interstitial cystitis, gastrointestinal disorders such as reflux oesophagitis and colitis, and even cancer. For example, TRPM8 is expressed in prostate cancer cells and therefore may have utility as a diagnostic marker122 and may even offer opportunities for therapy. The next few years should witness the direct testing in the clinic of the role of the different TRP channels in many of these indications. These results will help further define both the biology of TRP channels in sensory systems and highlight the most promising clinical indications for TRP channel blockers.
Potent TRPV1 agonists produce pain but also desensitize the receptor123. This inactivation reduces sensitivity to heat and other ligands, which can be used, with limited efficacy, to reduce pain124 — albeit with the problem of the pain that is produced before the desensitization (FIG. 4). This approach is being tested to treat several bladder conditions such as interstitial cystitis. Furthermore, agonists such as capsaicin and resiniferatoxin can lead to such a large calcium influx that they can produce degeneration of nociceptor axons at the site of application (topically) into joints or onto nerves. They may even cause a loss of the sensory neuron itself by calcium-mediated mitochondrial damage and cytochrome c release, leading to apoptosis125,126 when exposed close to the cell body (FIG. 4). High-dose topical capsaicin-induced peripheral terminal axonal loss does produce long-lasting analgesia in patients with HIV, but at the cost of intense initial pain in some patients127 and with the potential problems inherent in any neuroablative strategy128. Resiniferatoxin is more potent than capsaicin and can selectively ablate nociceptors when delivered intrathecally, which may have special utility for uncontrolled pain in a palliative setting129. Activation of TRPM8 by icilin is reported to produce analgesia by activating central inhibitory pathways50,105, and cold-induced analgesia is attenuated in TRPM8-deficient mice50,105 (FIG. 4).
TRP channels are non-selective cation channels with pores that seem to be large enough to allow cationic molecules with molecular masses of ~500 daltons to enter the cell130. TRPV1 shows a time-dependent and agonist-dependent increase in permeability to large cations owing to changes in the TRPV1 selectivity filter that are PKC-dependent and reflect increases in pore diameter131. This feature of the channels can be exploited to target small membrane-impermeable cationic drugs into cells that only express a particular TRP channel. Indeed, this approach can be used to produce analgesia. A quaternary derivative of the local anaesthetic lidocaine — QX-314 — that is ineffective in blocking sodium channels when administered extracellularly, because it cannot gain access to the inner face of the channels, inhibits sodium currents only in TRPV1-expressing nociceptors. This produces a dense (complete block of response to noxious stimuli), long-lasting local analgesia when co-administered with capsaicin, without impairing motor function or innocuous tactile sensitivity in the way that local anaesthetics do132. This approach produces local analgesia when injected subcutaneously, close to a peripheral nerve, and intrathecally. Because QX-314 is trapped in the nerve fibres the duration of effect is about ten-times longer than lidocaine and offers, therefore, a new approach to manage surgical and post-operative pain as well as labour pain. Because this strategy produces regional or local analgesia instead of a non-specific local anaesthesia with blockade of all sensory, motor and automatic fibres, patients will be able to mobilize earlier with less risk of hypotension due to autonomic blockade. To successfully use the technology in patients a non-pungent activator of TRPV1 would be ideal to minimize patient discomfort until sodium-channel blockade is achieved on entry of the charged membrane-impermeant sodium-channel blocker. Although this strategy uses a TRPV1 agonist, it is only as a means of targeting delivery of a drug to particular cells and not as an intrinsic analgesic as when used by itself. In principle, the technology should work for any TRP channels for which pore size is sufficient to allow entry of small cationic drugs, and for which expression is limited to nociceptors, such as TRPA1.
It is often difficult to obtain exact information regarding the status of molecules in early stages of clinical development. However, through a review of patent literature and company press releases we have noted significant activity, in particular with TRPV1 modulators (FIG. 5). Several companies, including GlaxoSmithKline, Amgen, Merck/Neurogen Astra Zeneca, Evotec/Pfizer and Glenmark, are conducting or have now completed Phase I studies with oral TRPV1 antagonists and some have progressed into Phase II for pain and/or migraine indications133. Other companies are exploring local application of TRPV1 agonists, including NeurogesX for neuropathic pain and Anesiva for post-surgical pain (see Further information for data from individual companies). There is considerable preclinical activity, not only with TRPV1 antagonists but also with TRPA1, TRPV3 and TRPV4 antagonists. Of note are two TRPV3 antagonists from Glenmark, GRC 15133 and GRC 17173, which demonstrated a significant reduction in thermal and mechanical hyperalgesia in models of neuropathic and inflammatory pain (see Further information).
The discovery of novel analgesics relies heavily on behavioural observations in animal models of pain before their testing in Phase II proof-of-concept clinical trials. However, there are concerns about the predictive value of animal models in pain. It is well known that most drug candidates fail in early clinical development phases, and poor translation from the pre-clinical to clinical setting is a likely reason for this low success rate. How can we improve the success ratio in the clinic? Perhaps of most importance is to ensure that when novel mechanisms such as TRP channel modulators are tested, the target mechanism is fully engaged by the drug molecule. We have postulated, for example, that a central site of action may be required in order to achieve maximum therapeutic benefit for TRPV1 modulators in certain inflammatory and neuropathic pain states. Target occupancy studies such as positron emission tomography (PET) scans can be used to demonstrate interactions between a drug and the relevant sites of action in the nervous system. PET studies could be performed as part of the Phase I studies, and will also provide valuable information on receptor distribution, receptor density and endogenous neurotransmitter release. The level of target occupancy can then be compared with pharmacological effects.
Over the past few years there has been significant progress in human experimental pain research, and new techniques have been developed for the induction and measurement of pain134. Thermal, mechanical, electrical and chemical stimuli can be applied to different tissues and then the responses recorded by a combination of imaging technology such as functional magnetic resonance imaging and PET, electrophysiological recordings such as electroencephalography and evoked potentials, and behavioural measurements135,136. Human experimental medicine studies were used to assess the pharmacodynamic and antihyperalgesic properties of the TRPV1 antagonist SB-705498 (REF. 133). To date, this is the only published TRPV1 clinical study with efficacy data; however, there are several other TRPV1 antagonists in early development (FIG. 5) and data from these studies are eagerly anticipated. The experimental protocol measured heat-evoked pain and skin sensitization, which were induced by either capsaicin or ultraviolet B irradiation. In comparison with placebo, SB-705498 was shown to reduce the area of capsaicin-evoked flare and to increase pain tolerance at the site of ultraviolet B irradiation. Both of these pharmacodynamic effects correlated with plasma exposure of SB-705498. This profile, which compared well with preclinical animal model studies, allowed this molecule to enter Phase II trials with increased confidence135. A further study has shown that PGE2 potentiates nociceptor activation by protons in human skin and that this is caused by TRPV1 sensitization137. This and other translational human experimental pain models will help the transition from preclinical to clinical pain studies for TRP channel modulator drugs.
TRPV1 seems like a highly exciting clinical opportunity based on early efficacy readouts, but potential side effects should be considered. TRPV1-expressing primary afferent neurons project to cardiovascular and renal tissues where they surround blood vessels in vascular beds138. A subset of these neurons release the neuropeptides CGRP and substance P following TRPV1 activation. These neuropeptides are potent vasodilators and natriuretic/diuretic factors, and activation of cardiac afferents that contain CGRP and substance P may be associated with cardio-protective effects138. TRPV1 antagonists may potentially produce unwanted cardiovascular side effects, although there is no data as yet from the early clinical trials that this will be an issue. Recent studies with the TRPV1 antagonist AMG-517 have demonstrated a long-lasting hyperthermia, with core body temperature rising to as much as 40 °C, in patients undergoing a dental pain clinical trial139.This effect seemed to be dose-dependent with some individuals more susceptible than others, and unfortunately the efficacy group of this trial could not be completed owing to this significant adverse effect. The magnitude of this effect seemed to be larger in the patient population than in healthy volunteers, suggesting a correlation with tonically active TRPV1 receptors. Trpv1−/− mice show an impaired response to hyperthermia, further implicating these ion channels in thermosensitivity and core body temperature regulation140. Although TRPV1 is expressed in thermosenstive neurons in the hypothalamus140 TRPV1 antagonists that are peripherally restricted still produce hyperthermia93, which implies that the site of action is outside the blood– brain barrier, and may be due to a blockade of tonically active TRPV1 receptors in abdominal viscera92.
At least two lines of evidence suggest this is not a potential ‘showstopper’ for TRPV1 antagonists. The first involves a study with AMG-8562, a TRPV1 antagonist that is selective for capsaicin and pH activation, but does not block heat-evoked activation of the receptor. AMG-8562 does not elicit a hyperthermic response in preclinical species at doses that are effective in rodent models of inflammatory pain141. Encouragingly, advances in ion-channel screening and the emergence of robust high-throughput electrophysiology systems open up the potential for identifying high quality, modality-specific TRP-channel blockers of this type. The second line of evidence comes from the study on SB-705498 in which hyperthermia was not observed in humans given 400 mg of this TRPV1 antagonist133. Given that this molecule antagonizes all modalities of TRPV1 activation, there may be further compensatory mechanisms that under-pin this apparent lack of hyperthermia. However, it is clear that there are significant challenges for drug companies to develop safe and effective medicines that target TRPV1. For example, clinical development of GRC 6211 — a TRPV1 antagonist that was in Phase II trials for osteoarthritis — has been suspended, with no indication, so far, as to why (Glenmark press release; see Further information).
The discovery that the capacity to detect many forms of noxious stimuli is mediated by high-threshold TRP channels expressed selectively in nociceptors has provided insights into the molecular basis of sensory labelled lines — neural input channels devoted to specific types of stimuli and in the peripheral nervous system. It has also provided an opportunity for the development of analgesics that are different from the non-steroidal anti-inflammatory drugs, opiates, anticonvulsants and antidepressants currently used.
The field has moved relatively quickly, from the cloning of TRPV1 to clinical trials within 10 years. Studies on TRPV1 raise the intriguing possibility that these ion channels have functional roles greater than initially anticipated, which may provide problems and opportunities. For example, while TRPV1 is an acute noxious heat receptor, it also seems to be a major player in various pain states such as bone cancer pain and neuropathic pain, as well as in thermoregulation and hippocampal plasticity. TRPA1 also seems to be a major player is sensing damage, and has great clinical potential as an analgesic and perhaps even more so for several respiratory diseases.
However, more studies are required to identify which TRP channels will be appropriate targets for which specific pain and other indications. Armed with genetic deletions for each of these ion channels, and specific small-molecule agonists and antagonists against many of them, the important questions should be addressed in the next few years.
We are poised to discover whether drugs that act on sensory TRP channels provide greater efficacy and fewer adverse effects in the treatment of pain than the currently available analgesics. In addition to targeting TRP channels with agonists to activate, desensitize and ablate nociceptors, and with antagonists to reduce TRP activity in the periphery and CNS, we can also use TRP channels as a drug delivery system to target, with TRP agonists, small cationic compounds selectively into those specific cells that express these ion channels. It seems likely that somehow, from these multiple and diverse therapeutic opportunities, TRP channels will provide a new means of reducing pain.
Competing interests statement
The authors declare competing financial interests: see web version for details.