The triple function of the capsaicin-sensitive sensory neurons: In memoriam János Szolcsányi

ABSTRACT

This paper is dedicated to the memory of János Szolcsányi (1938–2018), an outstanding Hungarian scientist. Among analgesics that act on pain receptors, he identified capsaicin as a selective lead molecule. He studied the application of capsaicin and revealed several physiological (pain, thermoregulation) and pathophysiological (inflammation, gastric ulcer) mechanisms. He discovered a new neuroregulatory system without sensory efferent reflex and investigated its pharmacology. The authors of this review are his former Ph.D. students who carried out their doctoral work in Szolcsányi’s laboratory between 1985 and 2010 and report on the scientific results obtained under his guidance. His research group provided evidence for the triple function of the peptidergic capsaicin-sensitive sensory neurons including classical afferent function, local efferent responses, and remote, hormone-like anti-inflammatory, and antinociceptive actions. They also proposed somatostatin receptor type 4 as a promising drug target for the treatment of pain and inflammation. They revealed that neonatal capsaicin treatment caused no acute neuronal death but instead long-lasting selective ultrastructural and functional changes in B-type sensory neurons, similar to adult treatment. They described that lipid raft disruption diminished the agonist-induced channel opening of the TRPV1, TRPA1, and TRPM8 receptors in native sensory neurons. Szolcsányi’s group has developed new devices for noxious heat threshold measurement: an increasing temperature hot plate and water bath. This novel approach proved suitable for assessing the thermal antinociceptive effects of analgesics as well as for analyzing peripheral mechanisms of thermonociception.

Introduction

János Szolcsányi (Budapest, February24 1938– Pécs, November 5 2018) was a Hungarian physician, pharmacologist, university professor, and full member of the Hungarian Academy of Sciences (Figure 1). He has studied the pharmacology of analgesics and anti-inflammatory drugs as well as the regulatory role of the capsaicin-sensitive sensory neurons.

Figure 1.

Prof. János Szolcsányi participating in the annual meeting of the European Neuropeptide Club in April 2007, Santorini, Greece (photo by Erika Pintér).

He worked at the Department of Pharmacology of the Medical School of the University of Pécs, Hungary, from 1970 until his death in 2018. He was the founder of the scientific workshop at the University of Pécs which is referred to as the “Szolcsányi School.” The sections of the present review have been written from the pens of his former coworkers and Ph.D. students about the most important scientific results they achieved under Professor Szolcsányi’s guidance.

If we had to describe Professor Szolcsányi in one sentence, we would say: “He was a creative, highly motivated, and optimistic scientist, who enjoyed the discovery itself, who was able to search for the solution of problems to the very end, and he usually found it … ”

The authors of the present review, Erika Pintér in 1985, Gábor Pethő in 1987, Zsuzsanna Helyes in 1995, Éva Szőke in 1996, Kata Bölcskei in 2000 joined Szolcsányi’s research group. All of them have profited enormously from this collaboration and are all prominent members of the Department of Pharmacology and Pharmacotherapy at the University of Pécs, Hungary. The first section summarizes the milestones of the discovery of the systemic anti-inflammatory action of the capsaicin-sensitive sensory neurons. The second part discusses evidence for somatostatin receptor type 4 (SST₄) as a promising novel drug target for the treatment of pain and inflammation. In the third section, a reevaluation of the “neurotoxic” effect of the neonatal capsaicin treatment on the sensory neurons is presented. Subsequently, the role of lipid raft disruption on the agonist-induced opening of the TRPV1, TRPA1, and TRPM8 cation channels is described. Finally, an account of the development of novel equipment for measuring the behavioral noxious heat threshold is provided, and the diverse utilization of this approach for studying the peripheral mechanisms of thermonociception is described. The effects of capsaicin on thermoregulation also form a significant part of Szolcsányi’s scientific legacy. As he published a comprehensive review on this topic in 2015 [1], these aspects are not dealt with in the present paper.

Discovery of the systemic anti-inflammatory action of the capsaicin-sensitive sensory nerves

Peripheral nerves usually contain a mixture of motor, sensory and autonomic fibers. If only the sensory fibers are to be stimulated, it is possible to block motor and autonomic nerve regulation by pharmacological tools, peripheral muscle relaxants, sympatholytics, and parasympatholytics. However, selective stimulation of the sensory nerves can be produced by electrical stimulation of the peripheral stump of the transected spinal dorsal root. In this case, the electrical impulses propagate in an antidromic direction toward the periphery to the nerve endings, where they induce the exocytosis of neuropeptide-containing vesicles, and the released peptides act on their own receptors, causing vasodilatation and plasma extravasation, i.e. the phenomenon of neurogenic inflammation is produced as a local effect. At the beginning, we mapped the capsaicin-sensitive innervation of the exteroceptive and interoceptive lumbosacral areas of the rat by antidromic electrical stimulation of the transected dorsal spinal roots. Prior to stimulation, animals received a plasma albumin binding dye, Evans blue, which extravasates with albumin and causes the inflamed area to become blue. The blued tissues showed a segmental distribution, and after extraction of the dye, the extent of inflammation could be determined by spectrophotometry. When the animals were systemically pretreated with capsaicin, no plasma extravasation occurred, demonstrating that the phenomenon is related to the capsaicin-sensitive system [2,3]. Thus, we have provided functional evidence for the existence of capsaicin-sensitive sensory fibers mediated efferent function and mapped the capsaicin-sensitive innervation of pelvic organs and lower body skin areas.

During this systematic analysis, we observed the phenomenon that electrical stimulation of unilateral lumbosacral dorsal roots causes intense local plasma extravasation in the skin and pelvic organs according to the innervation area on the same side. If a few minutes after the end of stimulation, a second inflammation is sought to be induced in another area of the body, for example, on the opposite leg, the second inflammation is always much smaller in magnitude compared to the initial response [2,4,5] (Figure 2). This surprising observation, first published as a short communication in 1988 in the journal “Agents and Actions” [2], proved to be serendipitous and formed the basis of our research for decades to come. Over the years, it has become clear that this original discovery not only has basic research significance but could also serve as the basis for developing new types of analgesics and anti-inflammatory drugs.

Figure 2.

After intravenous administration of the Evans blue dye, peripheral stumps of the cut L4–L5 spinal dorsal roots were electrically stimulated on one side and the other side (left panel). Plasma protein extravasation was detected in the skin of both hind limbs and quantified by spectrophotometric determination of Evans blue accumulation. As shown in the right panel, a 5 min delay (dt = 5 min) between stimulations resulted in about 50% reduced effect from the second stimulation. However, either a 60 min delay (dt = 60 min) between stimulations or simultaneous stimulation (dt = 0 min) results in no difference between the plasma extravasation responses between the two sides (source: Ph.D. Thesis by Erika Pintér, 1996).

These results suggested that during the development of primary local neurogenic inflammation, anti-inflammatory mediator(s) are released and transported by the bloodstream to distant sites in the body and inhibit the development of a second inflammation. The effect of the released anti-inflammatory agent(s) is no longer detectable after 60 min. Since the anti-inflammatory effect is also produced in adrenalectomized animals, cortical steroids do not play a role in the process [4].

Antidromic dorsal root stimulation also inhibited conjunctival neurogenic plasma extravasation induced by intraocular instillation of 0.1% capsaicin solution by about 34%. The primary reaction not only exerted an inhibitory effect on inflammation of purely neurogenic origin, but also significantly reduced the reaction induced by intraplantar carrageenin injection (1%) of mixed-type (containing both neurogenic and non-neurogenic components) [4]. The anti-inflammatory effect was found to be related to capsaicin-sensitive neurons for the following reason. When the first dorsal root stimulation was performed at the site where the sciatic nerve had been previously pretreated perineurally with 2% capsaicin solution, which desensitized the capsaicin-sensitive fibers, no inflammation was produced in the plantar skin on the same side, whereas the second stimulation on the opposite side caused plasma extravasation of the same intensity as the first stimulation. No systemic blood pressure changes were observed during the stimulations, thus ruling out the hypothesis that the anti-inflammatory effect was due to blood pressure and microcirculatory changes during stimulation [4]. After these very exciting observations, intensive research was started in our laboratory to answer the following questions:

1. What are the putative anti-inflammatory mediators?

2. Where is the anti-inflammatory substance released from? From the inflamed tissue or directly from the capsaicin-sensitive nerve endings?

3. What type of inflammatory processes or which phase of inflammation can it inhibit?

4. What modulatory effects are exerted by proinflammatory, and anti-inflammatory mediators released from capsaicin-sensitive nerve endings?

We hypothesized that they might be neuropeptides with anti-inflammatory effects. It is known from the literature that capsaicin-sensitive sensory neurons contain somatostatin (SOM), galanin, and opioid peptides, which, according to in vitro and some in vivo data, have anti-inflammatory effects. The main objective of the next phase of our experiments was to identify the endogenous anti-inflammatory agent. Based on the literature, it was suspected [6] that the substance we were looking for was SOM, and we sought evidence to detect this peptide and confirm its anti-inflammatory role.

Based on the results of our further studies, it has become clear that antidromic electrical or chemical stimulation of capsaicin-sensitive neurons (with mustard oil or capsaicin) releases anti-inflammatory mediators from the nerve endings, which are transported by the bloodstream to distant parts of the body, where they inhibit the development of a subsequent inflammation. The anti-inflammatory effect develops after 5–10 min, is also produced in adrenalectomized animals, and therefore cannot be the result of mobilization of adrenal cortical steroids [7]. However, the anti-inflammatory effect does not develop after systemic or local (perineural) capsaicin pretreatment. Also, no anti-inflammatory effect was observed when primary inflammation was induced via a purely non-neurogenic route by administration of dextran solution s.c. after chronic denervation. The hindlegs were denervated 5 days prior to dextran injection to induce degeneration of the nerve fibers and to exclude the neurogenic component of the inflammation [8]. Dextran induces mast cell degranulation [9], releases inflammatory mediators from mast cells including prostanoids, leukotrienes, platelet activation factor, cytokines, etc. Dextran treatment did not evoke anti-inflammatory action suggesting that the mediators released during the early phase of inflammation are not responsible for the systemic inhibitory effect. The fact that stimulation at a very low frequency (0.1 Hz) did not cause inflammation in the innervation area itself, but produced an anti-inflammatory effect, suggests that the inhibitory substance is released from the nerve ending rather than from the inflamed tissue. The optimum frequency of its release is lower than the excitation frequency required for the release of proinflammatory peptides (calcitonin gene-related peptide [CGRP], tachykinins) [8,10]. Previous studies have shown that microneurostimulation of cutaneous polymodal nociceptive fibers at frequencies below 0.5–1 Hz does not induce pain sensations [11,12]. It is assumed that inhibitory mediators released at very low stimulation frequencies may have an important regulatory role, being part of a rapidly activated endogenous neurogenic defense system with anti-inflammatory effects. Since mediators are released from sensory fibers, are transported by the bloodstream to distant areas of the body, and exert a systemic effect, this function can be termed “sensocrine” effect [5].

Although we cannot exclude the role of other anti-inflammatory neuropeptides, opioids or galanin [13,14], previous literature suggests that the observed phenomenon is due to the release of neuronal SOM [15–17], and we have focused on this peptide in the following experimental setups.

The evidence in favor of SOM mediation is as follows:

1. Exogenous SOM can also inhibit mustard oil-induced plasma extravasation [18].

2. SOM antiserum pretreatment prevents the anti-inflammatory effects of antidromic and chemical neurogenic inflammation [8,10].

3. The anti-inflammatory phenomenon does not occur even after cysteamine pretreatment causing SOM depletion [8,10].

4. Plasma levels of SOM increase severalfold after mustard oil or antidromic sciatic nerve stimulation [8,10].

It is legitimate to ask why neuronal SOM has a systemic effect, whereas proinflammatory peptides (tachykinins, CGRP), also present in the capsaicin-sensitive nerve endings, have a predominantly local effect. After mustard oil smearing, plasma levels of substance P (SP) were also measured by radioimmunoassay, but mustard oil stimulation did not cause significant changes in plasma levels of SP (unpublished observation by our group). In 1999, Dux et al. described, based on immunohistochemistry studies, that capsaicin-sensitive cutaneous afferent fibers in the dermis were rich in SP and CGRP, whereas SOM-containing fibers could only be detected in the subepidermal layer [19]. It is assumed that SOM released subepidermally reaches the vessels and then the systemic circulation more easily, whereas proinflammatory peptides released in the epidermis can cause local effects. Of course, the distribution of receptors of the released neuropeptides (SST, NK₁, CGRP₁) can also have a major influence.

Our investigations have also presented a large body of evidence demonstrating the antinociceptive effects of SOM as well. The anti-inflammatory/pain-relieving SOM theory, which has been evidenced by in vivo animal experiments, could provide a rational explanation for the therapeutic analgesic efficacy of acupuncture, transcutaneous electrical stimulation, and counter-irritant treatments. These procedures can be used to stimulate sensory nerve endings with subsequent release of SOM which gets into the systemic circulation and exerts systemic analgesic and anti-inflammatory effects [20].

The basic physiological-pharmacological research forms the basis of applied drug development, and the main goal of a pharmacologist is to develop drugs that are better than existing ones in terms of efficacy, ease of use, and side effect profile. Therefore, we had the opportunity, based on our own experimental observations, to launch into the field of applied drug discovery. The aim was to develop a compound acting on the SST receptors that is a potent and effective anti-inflammatory and analgesic, with the fewest possible side effects, well tolerated, stable, and easy to use by the patients.

Although SOM injection is used in certain acute cases, this agent is not suitable for the role mentioned above. When given by infusion, SOM has been shown to be an effective agent for the treatment of excessive secretion from endocrine tumors of the gastrointestinal tract and for the treatment of acute, severe gastrointestinal bleeding. There are also stable octapeptide analogs on the market, the best known of which is octreotide (Sandostatin) used for the symptomatic treatment of endocrine tumors [21]. Analysis of the function of SOM receptor types (SST_1–5) revealed that the endocrine effects are mediated by members of the somatotropin release inhibiting factor (SRIF) SRIF1 group of receptors (SST₂, SST₃ and SST₅), whereas the anti-inflammatory and analgesic effects are mediated by members of the SRIF2 group (SST₁ and SST₄). As SOM analogs are also considered antitumor agents due to their antiproliferative activity, a number of stable analogs have been synthesized [20].

The SST₄ somatostatin receptor as a promising novel drug target for the treatment of pain and inflammation

We were the first to demonstrate that the antinociceptive and anti-inflammatory effects of SOM are mediated predominantly by the SST₄ receptor [18,20,22,23]. At the cellular level, the antinociceptive effect is related to the stimulation of K⁺ channels and inhibition of voltage-gated Ca²⁺ channels through G_i protein-coupled SST receptors via reduction in intracellular cAMP levels through inhibiting adenylate cyclase. These mechanisms consequently hyperpolarize the neuron and reduce transmitter exocytosis [24].

Although cloning of these receptors has facilitated research on the physiological/pathophysiological role of SOM, important progress has been achieved by generating gene-deficient mice. The therapeutic use of native SOM is hampered by its broad spectrum of action and very short (less than 3 min) plasma elimination half-life [25]. However, stable, selective SST₄ agonists may offer novel analgesic and anti-inflammatory options without endocrine side effects. The stable cyclic heptapeptide TT-232 (D-Phe-Cys-Tyr-D-Trp-Lys-Cys-Thr-NH2), synthesized by the Peptide Biochemistry Research Group of the Hungarian Academy of Sciences, Budapest (György Kéri and his colleagues), binds with high affinity to SST₄ [26]. The aromatic side chain of D-phenylalanine at the N-terminal end results in a conformation that protects the disulfide bridge from enzymatic degradation. TT-232 did not affect growth hormone and gastrin secretion but had potent antiproliferative effects and exerted significant inhibitory effects in acute nociception and inflammation models [18,27,28]. TT-232 was found to be a toxic over a wide dose range (up to 5 mg/kg), therefore the LD₅₀ value was not applicable. In mice, repeated administration of the highest dose of 120 mg/kg for several weeks did not cause death, with only 10% showing a slight, transient loss of body weight. Hematological parameters, qualitative blood counts, histopathology of various organs and gastrointestinal mucosal integrity were unchanged after prolonged administration of this high dose of TT-232. This dose is approximately 1000-fold higher than that one causing a significant inhibitory effect in acute inflammation models [18,22]. In rats, whole-body autoradiographic studies detected very low concentrations of only 0.1 µg/g TT-232 in the brain 30 min after 2 mg/kg i.v. 14C-TT-232 injection, indicating minimal penetration across the blood–brain barrier. No central nervous system effects were observed after the highest dose of 5 mg/kg i.v. (unpublished data, available in the preclinical documentation of the compound).

The selective, high affinity SST₄ receptor agonist J-2156 is a non-peptide sulfonamidopeptidomimetic (1ʹS,2S)-4-amino-N-(1’-carbamoyl-2’-phenylethyl)-2-(4”-methyl-1”naphthalenesulfonamino)-butanamide. This compound was synthesized by Juvantia Pharma (Turku, Finland). J-2156 binds to the human SST₄ receptor with nanomolar affinity, which exceeds the binding affinity of native SOM, and shows nearly 400-fold selectivity for SST4 compared to the other SST receptors [29]. In a cAMP assay indicating receptor activation, it behaved as a full agonist similar to native SOM-14 or SOM-28. In another G-protein activation functional assay, it produced 2.5-fold stronger responses than native SOM. Based on these properties, this molecule was considered as a “superagonist” [29]. Further in vitro studies showed that J-2156 does not cause desensitization after repeated administration, an important consideration for the potential therapeutic use of this compound [30].

TT-232 effectively inhibited the release of sensory neuropeptides in vitro and acute neurogenic and non-neurogenic inflammatory processes in vivo both in rats and mice. The minimum effective dose was 10 µg/kg i.p. in mustard oil-induced neurogenic and dextran- or bradykinin-induced non-neurogenic acute inflammation models [18,22,26]. J-2156 inhibited electrical field stimulation-evoked neuropeptide release from the capsaicin-sensitive sensory nerve endings of isolated rat trachea in a concentration-dependent manner and significantly inhibited mustard oil-induced acute neurogenic plasma protein extravasation in the rat paw skin. Based on these results, it was proposed that the inhibition of the release of inflammatory sensory neuropeptides (SP, CGRP) from capsaicin-sensitive sensory nerves plays a role in the neurogenic inflammation-reducing effect. No dose–response relationship was observed in several in vivo experiments, mainly in models including a neurogenic inflammatory component. This could be explained by the fact that at higher doses, J-2156 may inhibit the release of the endogenous anti-inflammatory SOM from capsaicin-sensitive sensory nerves, in addition to inhibiting the release of proinflammatory neuropeptides. Considering that J-2156 was also able to inhibit non-neurogenic inflammation induced by dextran or bradykinin, its mechanism of action has other postjunctional components, probably also exerting an inhibitory effect directly on vascular endothelial and/or inflammatory cells, especially mast cells. J-2156 did not affect rat arterial blood pressure or heart rate even at a dose of 100 µg/kg [18,20,22,31].

Receptor binding results revealed that J-2156 has orders of magnitude higher affinity for SST₄ than TT-232 (K_d values: 1. 2 ± 0.4 nM vs. 200 ± 10 nM). TT-232 also binds to SST₁ receptors to a significant extent [K_d: 1300 ± 400 nM; ²⁶], and inhibits tyrosine kinases [28,32] which might also be involved in mediating its anti-inflammatory effects. The major advantage of J-2156 over TT-232 is undoubtedly its non-peptide structure and therefore, its oral administration. The anti-inflammatory effect of SST₄ stimulation on other types of inflammatory cells has been demonstrated in isolated peritoneal macrophage cultures.

J-2156 significantly inhibited interleukin-1β production by stimulated peritoneal macrophages with a maximum inhibitory effect of 50–70%, but no concentration–response relationship could be established in this experimental setup. This proinflammatory cytokine produced by a variety of cells plays a major role in the development of many inflammatory processes (e.g. in the respiratory tract and in joints).

The peripherally acting TT-232 had remarkable analgesic effects in various nociceptive models in both rats and mice. In conventional chemonociception assays, in a new model of thermonociception, and in models of traumatic mononeuropathy and diabetic polyneuropathy, it produced significant inhibitory effects at very low doses. TT-232 was found to be 1000 times more potent than diclofenac in the formalin test and about 300 times more potent than both morphine and diclofenac in the two thermonociception tests, compared to previous results [33]. The first phase of the formalin test is a consequence of direct stimulation of nociceptors, while the second is a consequence of gradually developing inflammation [34]. TT-232 reduced nocifensive behaviors in both phases, suggesting that it was able to reduce both nociceptor activation and sensitization produced by inflammation. Similarly, in thermonociception assays, it had inhibitory effects on both the noxious heat-induced withdrawal response of untreated animals and the thermal allodynia induced by the transient receptor potential vanilloid 1 (TRPV1) receptor agonist resiniferatoxin (RTX). TT232 might reduce nociceptor excitability and transmitter release due to G_i protein activation [35]. It is also known, however, that nerve growth factor (NGF), which is released upon tissue injury, can bind to the tyrosine kinase-coupled receptor (trkA) and sensitize capsaicin-sensitive nociceptors and induce hyperalgesia [36–38], so it cannot be excluded that tyrosine kinase inhibition also contributes to the antinociceptive effects. The anti-inflammatory effect of TT-232 was shown to be dose-dependent [18,22,39], but in nociceptive studies, a bell-shaped dose–response curve has been observed in several cases. J-2156 also showed potent and broad-spectrum analgesic effects. It inhibited acute nocifensive reactions in a formalin assay, as well as chronic inflammatory and neuropathic mechanical hyperalgesia at doses of 10 and 100 µg/kg. The maximum antinociceptive and antihyperalgesic effect was about 50–70%, which is significant but less than the maximum inhibition seen with TT-232. TT-232 was highly effective in the traumatic mononeuropathy model, with 10 and 20 µg/kg doses causing hypoalgesia. Another notable difference is that, unlike TT-232, J-2156 did not reduce nocifensive responses in phase 1 of the formalin test. Based on our results, the SST₄ receptor can be an excellent target for drug development as its agonists have a broad analgesic spectrum that includes difficult-to-treat neuropathic pain conditions.

We described SST₄ expression patterns in the nociceptive and antinociceptive pathways of both mouse and human brain and provided neurochemical and electrophysiological characterization of the SST₄-expressing neurons. Intense or moderate SST₄ expression was shown predominantly in glutamatergic neurons in the major components of the pain matrix also involved in mood regulation, such as prelimbic cortex, anterior cingulate cortex, amygdala, CA1 region of the hippocampus, medial habenula, and several limbic areas. J-2156 significantly decreased the firing rate of layer V pyramidal neurons by augmenting the depolarization-activated, non-inactivating K⁺ current (M current), leading to remarkable inhibition. These were the first translational results explaining the mechanisms of action of SST₄ agonists as novel analgesic and antidepressant candidates [40]. Furthermore, SST₄-deficient mice showed higher susceptibility to chronic variable stress pointing out that SST₄ activation is involved in stress-induced behavioral and neuroendocrine alterations with a crucial role in plastic changes in the amygdala [41]. Small non-peptide molecules with good brain penetration abilities are needed to induce antidepressant and anxiolytic actions besides the analgesic effect, which would be particularly valuable in chronic pain conditions with stress and/or mood disorders as common co-morbidities. Although the central nervous system penetration of J-2156 was not measured, central nervous system effects are likely based on the antidepressant-like and anxiolytic actions after its systemic injection [41]. To overcome potential species differences of SST4 expression and function between humans and mice, we generated a SST4 humanized mouse line to serve as a translational animal model for preclinical research. RNAscope in situ hybridization revealed the presence of hSST₄ transcripts in glutamatergic excitatory neurons in the CA1 and CA2 regions of the hippocampus; in the GABAergic interneurons in the granular layer of the olfactory bulb, and in both types of neurons in the primary somatosensory cortex, piriform cortex, prelimbic cortex and amygdala. This novel SST₄ humanized mouse line provides a unique opportunity to investigate human SST₄ receptor expression, function, and drug developmental potentials [42].

We provided strong proof of concept evidence that SOM exerts potent analgesic and anti-inflammatory actions via SST₄ located both in the periphery and the central nervous system. Therefore, SST₄ agonists are promising novel drug candidates for neuropathic pain and neurogenic inflammation but rational drug design has not been possible due to the lack of knowledge about the 3-dimensional structure of SST₄. We modeled the SST₄ structure, described its agonist binding properties, and characterized the binding of novel small molecule SST₄ agonists using an in-silico platform and SOM displacement in the competitive binding assay on SST₄-expressing cells. We defined high- and low-affinity binding pockets of SST₄ for our ligands, binding of the highest affinity compounds were similar to that of the reference ligand J-2156. Strong G protein activation was shown with the highest potency of 10 nM EC₅₀ value and highest efficacy of 342%. Oral administration of 100 μg/kg of several compounds significantly inhibited acute neurogenic plasma protein extravasation in the paw skin and diminished sciatic nerve ligation-induced neuropathic hyperalgesia [43–45]. Lead selection and optimization processes are currently ongoing to determine the most appropriate drug candidate for preclinical development.

Reevaluation of the “neurotoxic” effect of the neonatal capsaicin treatment on the sensory neurons

Szolcsányi’s research group reevaluated the “neurotoxic” effect of the neonatal capsaicin treatment in sensory neurons. Systemic administration of TRPV1 agonists capsaicin or RTX caused damage of small, dark B-type neurons in the rat trigeminal ganglion or dorsal root ganglion (DRG) in adult [46–49] and neonatal rats [47,50–55] and mice [55,56]. After neonatal capsaicin pretreatment in adult animals significantly fewer B-type sensory neurons and non-myelinated fibers could be seen [47,50–52,54,57]. Similar cell death after adult pretreatment could not be seen, and the number of dorsal root C-type afferent fibers was not reduced [46–48]. In neonatal rats, the B cell damage was described as an acute neurotoxic effect which manifested itself within half an hour, and capsaicin was regarded as a sensory neurotoxin [50,51,58], therefore the functional impairment was explained by the absence of neurons. After adult pretreatment with capsaicin or RTX, severe mitochondrial damage was observed in some B cells, which lasted for days or even weeks after injection, but neither decrease in the number of dorsal root axons nor acute death of the sensory neurons were found [46–48,59]. The explanation of the qualitative differences between the effect of neonatal and adult pretreatment was not known. In the development of pain-relieving drugs that act on the TRPV1 receptor, an important aspect is whether the lead molecules used for the synthesis of new compounds have a neurotoxic effect that causes acute cell death or not. However, the rapid cell death-inducing effect of neonatal capsaicin pretreatment has only been reported on the basis of some ultrastructural changes in few neurons. Interestingly, quantitative morphometric measurements after the treatment had not been carried out prior to our study.

Our electron microscopic studies and quantitative data demonstrated that neonatal capsaicin pretreatment induced slow mitochondrial damage and slow cell number loss. It was known from previous morphometric studies that neonatal treatment with 50 mg/kg capsaicin caused cell depletion in DRG neurons 3–6 months after injection [52,53,60]. The number of afferent fibers without myelin sheaths was also reduced [51,52,55,58,61]. Applying a higher dose of capsaicin (100 mg/kg), the number of myelin sheath fibers [57] and the number of A-type cells is also reduced [52,62]. However, only Otten and colleagues [63] reported cell counting data that already showed a decrease in cell number at DRGs two weeks after capsaicin treatment. No cell count determinations were made at an earlier time point. It had long been believed that the decrease in cell number induced by capsaicin was due to the death of B cells which appeared to be rapid necrosis within 30 min [51,58,64]. In mouse experiments, a slower degeneration of B cells was described 2–3 days after treatment [65]. The neurotoxic effect of capsaicin on sensory neurons could also be detected in vitro, but only at high concentrations and only after a longer latency period. When capsaicin was added to cell cultures prepared from DRG at a concentration of 30 mM, only about 1.9% of the cells died after 24 hours [66], even though this concentration is much higher than that resulting from 50 mg/kg subcutaneous in vivo dose. In samples taken from blood, brain and spinal cord, its amount reached a maximum of 0.4–0.5 mg/g 5 hours after treatment [67]. Sugimoto and coworkers described apoptotic neurons 12–48 hours after neonatal capsaicin treatment (50 mg/kg, s.c.) [47,68], but no signs of the previously described rapid neurotoxic cell death were found. No signs of apoptosis were observed during adult pretreatment. These facts did not support the hypothesis that capsaicin causes rapid, neurotoxic cell death. In our experiments, 24 hours after capsaicin pretreatment, apoptotic cells were not found, but osmiophilic, necrotic cells were present, nuclei and cell organelles were damaged, as previously described in rats [50,51] and mice [65,69]. It should be mentioned, however, that after subcutaneous injection of capsaicin, apnea, bradycardia, and hypotension occurred, so artificial ventilation was necessary for the survival of the animals. It can be assumed that, despite artificial ventilation, severe hypoxia developed, leading to apoptotic or necrotic cell death in the case of some cells [70]. Five days after the treatment, however, we did not find a significant decrease in the number of cells. Therefore, the reduction in the number of B cells after three weeks cannot be considered because of an early, acute cell death. Previous studies did not report a decrease in the number of cells that can definitely be attributed to acute cell death. The late-onset cell death is probably caused by the lack of NGF uptake occurring as a result of the degeneration of peripheral and central nerve endings in case of neonatal pretreatment. After treatment in adults, this effect is not accompanied by cell death. NGF is essential in the prenatal and early postnatal development of nociceptive afferents [71]. Depriving cells of NGF in newborn animals leads to cell death. Systemic capsaicin treatment in adult animals caused the destruction of central and peripheral terminals of axons [72–74]. Similar damage in newborn animals prevented the transport of NGF from axons to the cell body, so they may die. The decrease in the number of DRG cells after capsaicin treatment could be prevented by the administration of NGF [63]. Since NGF, in addition to being important for cell survival, also increases the sensitivity of cells to capsaicin [75,76], we did not give capsaicin and NGF simultaneously to rats during the experiments as Otten and coworkers, but we started NGF administration one day later. Thus, we did not influence the development of early ultrastructural changes caused by capsaicin. As a result of systemic capsaicin or RTX treatment, in parallel with the prolonged functional disturbances of the nociceptive primary afferents, the damage of mitochondria could also be observed, as previously described in adult animal [46,48,73,77]. A similar change did not occur in vitro in a calcium-free environment [78]. By opening the TRPV1 cation channel capsaicin increased the intracellular calcium ion concentration [79,80], the mitochondria became saturated with calcium leading to structural damage [81]. The question why this does not lead to immediate cell death or why it is not followed by restitution was unanswered. Such mitochondria could still be seen weeks after treatment. NGF treatment prevented the decrease in cell number but had no effect on the disorganization of mitochondria.

At this point, further studies could shed light on the role of mitochondria in the cell’s life cycle as well as in the molecular biological events triggered by TRPV1 activation. In summary, our research group could conclude that neonatal capsaicin treatment caused long-lasting selective ultrastructural and functional changes in B-type sensory neurons, similarly to adult treatment. Quantitative morphometric data could not support the early neurotoxic effect of capsaicin. A significant decrease in B-type neurons could only be detected after three weeks and could be prevented by NGF administration. NGF administration antagonized the antinociceptive effect induced by capsaicin administration but did not prevent the neurogenic anti-inflammatory effect induced by capsaicin administration and the mitochondrial swelling lasting for several weeks [82].

Investigation of the role of lipid environment in the function of TRP ion channels

Previous studies on TRP ion channels characterized important properties of these receptors but our knowledge about the potential function of the lipid rafts surrounding them in the plasma membrane is limited. After the description of lipid rafts, it has been reported that they have important functional significance for example around the nicotinic acetylcholine receptors or human 5-HT₇ receptors [83–87] as microdomains rich in cholesterol, sphingomyelin and gangliosides. In the early 2000s few studies were published with controversial results about the role of lipid rafts in the function of TRPV1 channels. Cholesterol depletion by methyl β-cyclodextrin (MCD) did not result in any change in heat-induced currents of TRPV1-transfected Xenopus laevis oocytes [88], while in DRG neurons the magnitude of capsaicin-evoked currents was significantly reduced [89]. In contrast, it has been reported previously that cholesterol depletion did not modulate ^3HRTX binding to TRPV1 receptors in rat C6 glioma cells [90]. Beyond cholesterol depletion other treatments are also suitable for disintegration of lipid rafts. Sphingomyelinase (SMase) depletes membrane sphingomyelin by hydrolysis of sphingomyelin to phosphocholine and ceramide [91,92]. The third mechanism for lipid raft disruption is to inhibit the glycosphingolipid synthesis by myriocin treatment [93,94]. The aim of our research group was to investigate whether the change in lipid rafts surrounding the TRP channels alters their chemical activation. We examined the rise in intracellular Ca²⁺ concentration in sensory neuronal culture, receptor-expressing cell line and performed fluorescence spectroscopy and filipin staining to examine the properties of plasma membrane of CHO cells after lipid raft disruptor treatments (Figure 3) [95]. We used different vanilloid and nonvanilloid activators of the TRPV1 channel and agonists of transient receptor potential ankyrin 1 (TRPA1) (allyl isothiocyanate, formaldehyde) under cholesterol, sphingomyelin or gangliosides depletion [95,96]. In transient receptor potential melastatin 8 (TRPM8) receptor activation the involvement of lipid rafts was proven, MCD treatment shifted the threshold for TRPM8 activation to a higher temperature [97]. TRPM8 and TRPM3 are also thermosensor TRP channels, menthol, icilin and temperatures below 26°C activate TRPM8 [98–101]. TRPM3 is expressed in a variety of neuronal and non-neuronal tissues, pregnenolone sulfate derived from cholesterol is a potent activator [102–105].

Figure 3.

A: Filipin staining of CHO cells on control and MCD-treated plates. MCD reduced the cholesterol-binding filipin labeling of the plasma membrane in CHO cells. Micrographs have been generated using an Olympus Fluoview-1000 system on an Olympus IX81 microscope stage equipped with an Olympus DP70 digital camera and through an Olympus UPlan FL N, Phase2 objective (40x/0.75). B: Effect of MCD on capsaicin-evoked Ca²⁺ accumulation in TRPV1-expressing CHO cell line in radioactive ⁴⁵Ca²⁺ uptake experiments. Ca²⁺ influx is presented in % of control (without MCD treatment). *P < 0.05, **P < 0.01 (MCD-treated vs. control; One-way ANOVA).

We demonstrated that depletion of cholesterol by MCD diminished the Ca²⁺ response evoked by capsaicin but not by RTX, anandamide or low pH in TRPV1-expressing cells. It is not surprising that lipid raft disruption failed to influence the proton-evoked activation in the plasma membrane. It has been described earlier on cell lines transfected with chimeric TRPV1 channels, amino acids located at the external part of the pore loop are responsible for this response [106,107]. Also, earlier structure–activity relationship studies explained the different effect of MCD treatment on Ca²⁺ responses evoked by RTX or capsaicin. Partly different binding sites are responsible for TRPV1 channel opening in the presence of these vanilloid agonists. The H-bonding ability of capsaicin is necessary for the action while it is not important in the case of RTX [108,109]. In our study the effect of RTX was not inhibited by cholesterol depletion on CHO cells, most likely the large hydrophobic skeleton of the diterpene with a 3α-keto substitution is sufficient for binding in these cells which is critical for RTX action [108,109]. In contrast, cholesterol depletion inhibited the actions of both RTX and capsaicin in native sensory neurons. Our data suggest that there is a difference between TRPV1-expressing cell line and native sensory neurons in the lipid raft surrounding the ion channel in plasma membrane and this may be partly the cause of differences in TRPV1 gating mechanisms induced by the two vanilloid agonists. In addition, in sensory neurons interaction between TRPV1 and other TRP channels should also be taken into consideration [110]. We described that lipid raft disruption through different pathways concentration-dependently and significantly diminished the agonist-induced channel opening of the TRPV1, TRPA1, TRPM8, but not the TRPM3, cation channel in native sensory neurons. The results of fluorescence spectroscopy as well as the disappearance of filipin staining from the plasma membrane of TRPV1 receptor-expressing CHO cells and sensory neurons provided evidence for the depletion of cholesterol after MCD treatment. To visualize at the cellular level the effect MCD pretreatment on cholesterol level of plasma membrane CHO cells were stained with the cholesterol-binding compound filipin, the filipin staining faded in CHO cells due to cholesterol depletion (Figure 3a). MCD, SMase and myriocin incubation had no effect on pregnenolone sulfate-induced TRPM3 activation on native sensory neurons in our investigation. We could explain this finding with the special features of the TRPM3 receptor which is also a cation-permeable ion channel like the other TRP channels but with several unique features [104,105]. Administration an MCD/cholesterol complex decreased human and mouse TRPM3 channel activation, while MCD treatment elevated the pregnenolone sulfate-evoked TRPM3 channel activity on contractile and proliferating phenotypes of mouse vascular smooth muscle cells in two previous studies [111,112]. It had been suggested by Drews and coworkers that TRPM3 is present in a quaternary complex with a presently unknown auxiliary protein, and this channel form is quite strong to resist the cholesterol depletion by MCD or SMase treatment [111].

All these in vitro data clearly demonstrated that lipid raft disruption decreases TRP channel activation, nonetheless there were only few reports describing this effect in vivo. MCD significantly decreased the hyperalgesia in the RTX-induced neuropathy model in mice [113] and after prostaglandin E₂ administration [114]. In our experiments, we investigated the effect of lipid raft disintegration in TRPV1 and TRPA1 activation-related mouse pain models. We also revealed in vivo evidence for the antinociceptive effect of lipid raft disruption by SMase or myriocin [91–93,115]. We also proved that SMase and myriocin decreased the RTX-induced thermal allodynia as well as capsaicin-evoked eye wiping. We described that SMase diminished the duration of the formaldehyde-evoked acute nocifensive behavior and the RTX-induced mechanical hyperalgesia.

Conformational changes of TRP channels in gating functions can be analyzed by genomics and proteomics. Much less straightforward methodical possibilities are available to reveal the role of lipid rafts around membrane proteins in living cell plasma membranes. Since TRP ion channels operate under physiological conditions in the nerve endings, it is very difficult to prove the role of lipid rafts in the gating of ion channels that induce nociception. We suggest that disruption of lipid rafts changes the position of the ion channels in the membrane, some binding sites will not be accessible to agonists and it may affect the gating of the ion channels.

Our previous in vitro data clearly demonstrated the ability of MCD, SMase and myriocin to significantly and concentration-dependently inhibit TRPV1, TRPA1 and TRPM8 ion channel activation both on receptor-expressing cell line and primary cultures of sensory neurons [95,96]. These data are well supported by novel in vivo results [115]. Based on the in vitro findings, we can conclude that the hydrophobic interactions of ligands at the TRP channel/lipid raft interface play a significant role in drug action. Our in vivo results demonstrate that membrane lipid modification might open novel analgesic opportunities.

Development of novel equipment for measuring the behavioral noxious heat threshold and utilization of this approach for studying peripheral mechanisms of thermonociception

Conventional methods for studying thermonociception (i.e. heat-evoked “pain” in animals) are based on the following common principle. The tail or the paw(s) of a rat or mouse is exposed to noxious radiant heat or to a painfully hot water or metallic surface and the latency time, i.e. time elapsing until occurrence of an avoidance (withdrawal) reaction is recorded [for rev. see 116]. These methods, including the hot plate, tail-flick, tail immersion, paw immersion and plantar test, are widely used albeit they have some disadvantages. On the one hand, the latency time depends on both the actual skin temperature and the intensity of heat stimulation. On the other hand, latency as time is not comparable with the heat threshold for activation (measured as a temperature) routinely measured in single-fiber recording and patch clamp experiments as well as in human studies. János Szolcsányi was driven to develop a device for measuring the noxious heat threshold (i.e. the lowest hot temperature that evokes a nocifensive reaction) by an apparent contradiction. The sensory desensitizing effect of high concentrations of capsaicin was clearly demonstrated in single-fiber recording experiments as an elevation of the activation heat threshold of the units [117], however, it could not be unequivocally demonstrated in thermonociceptive tests measuring latency times [118–123]. As a way out, Szolcsányi hypothesized that the in vivo tests available are inappropriate, and he developed a very simple home-made device to measure an approximate heat threshold: a hindpaw of the rat or the tail of the mouse was successively immersed into a series of water tanks having 1°C difference in temperature until a heat-induced withdrawal reaction occurred [117,123]. The actual water temperature was considered the noxious heat threshold. With this method, the heat-desensitizing effect of capsaicin could be clearly demonstrated.

More than 20 years ago, we developed an increasing-temperature hot plate with the cooperation of an industrial partner (Supertech Ltd, Pécs, Hungary) [33]. The equipment consisted of three parts: a heating surface (20 × 15 cm metal plate) with the heating unit below and a Plexiglas observation chamber above, a control unit and a PC. A special software allowed the heating of the metal plate by selecting different heating rates. The cutoff temperature was 50°C. During the heat threshold measurement, the rat or the mouse was placed in the observation chamber above the metal plate of the apparatus having an initial plate temperature of 30°C. The plate was then heated at a steady rate of 6 or 12°C/min using a computer control until a nocifensive reaction occurred in any of the animal’s limbs. The typical response was licking of a hindleg, shaking or lifting of the limb, jumping up of the animal was rarely observed. The plate temperature that elicited any of these responses in any leg was considered the noxious heat threshold. The animal was then immediately removed from the chamber, the heating was interrupted, and the plate was cooled to below 30°C by placing an ice-cold steel cover over it. The threshold measurement was repeated after 30 min, and the average of the two thresholds was taken as the baseline (control) noxious heat threshold.

Subsequently, another piece of equipment has also been developed in cooperation with Experimetria Ltd [Budapest, Hungary; ¹²⁴] (Figure 4). The increasing-temperature water bath consists of two parts: a plastic cylinder (12 cm inner diameter, 14 cm height) with an electric heating unit at the bottom and a control unit to heat the water in the cylinder quickly and relatively evenly. The control unit indicates the current temperature of the water bath which is measured by a thermocouple 35 mm below the surface of the water. The heater can be set to an initial heating temperature (30 or 40°C) and a heating rate (6, 12 or 24°C/min). The heating can be stopped by a foot switch, in which case the current water temperature is fixed on the display. The water is cooled back by a pump that draws cold tap water into the water bath, from where the overflow is drained. For the measurement of the baseline heat threshold, an initial temperature of 30°C, a heating rate of 24°C/min and a cutoff temperature of 53°C were set. One hindpaw of the rat gently held in the vertical body position was immersed in the water bath to a depth of about 3–4 cm. The heating was then turned on. When the animal pulled its leg out of the water, the heating was immediately turned off and the temperature read by the apparatus was taken as the noxious heat threshold of the paw. The threshold measurement was repeated 30 min later on the same limb and the average of the two thresholds was taken as the baseline (control) noxious heat threshold. When using mice, the animal was held in an upright position with a gentle grip over the water bath so that the tail of the animal reached a depth of about 2 cm into the water. The initial temperature was 40°C and the heating rate was 24°C/min. The cutoff temperature was set at 53°C. The typical end point was a tail withdrawal or shake. When either of these occurred, the animal’s tail was removed from the water bath, the heating was turned off, and the current water temperature was taken as the nociceptive heat threshold.

Figure 4.

Measurement of the noxious heat threshold of the rat hindpaw by the increasing-temperature water bath (photos by Kata Bölcskei).

As a methodological improvement, we have developed an arrangement where the mouse does not need to be hand-held during the measurement of the heat threshold of the tail as the animal is placed 10 min before the measurement in a plastic cylinder with a top-closing lid from which the tail protrudes downwards, and holes on the side of the cylinder allow free breathing [125]. This arrangement both allows a longer adaptation period and eliminates a potential source of error due to restraint. The cylinders were hung in a vertical position on a stand on which they could slide horizontally. During the threshold measurement, the next cylinder in line was positioned above the water bath at a height such that the animal’s tail hanging down was about 4 cm into the water. The water was then heated up from a starting temperature of 30°C at a rate of 24°C/min. The heating continued until the animal pulled its tail out of the water. The animal was then kept in the cylinder for a maximum of 90 min until the end of the experimental protocol. The experiment continued with the next animal’s thermal threshold measurement during which the next cylinder was placed over the water bath, etc.

As part of the biological validation of the novel devices, the reproducibility of the noxious heat threshold measurement was examined [33]. The control heat threshold of the hind paw of the rat measured by the increasing-temperature hot plate was 45.3°C. Upon repeated measurements at intervals of either 5 min, 30 min or 24 hours, no significant alteration of the threshold was detected. The reproducibility of the heat threshold measured by the increasing-temperature water bath was comparably good [for details see 124]. As a next step, the effects of reference analgesics on the noxious heat threshold were assessed [33]. Morphine, diclofenac or paracetamol (acetaminophen) given i.p. dose-dependently raised the noxious heat threshold measured with the increasing-temperature hot plate i.e. they exerted a thermal antinociceptive effect. The minimum effective doses (MEDs) were 3, 10 and 200 mg/kg, respectively. These values are similar (morphine) or even lower (diclofenac, paracetamol) than those obtained in conventional thermonociceptive tests employing latency time measurement.

Our starting hypothesis was that any agent or other stimulus causing a shortening of latency time until the onset of a withdrawal reaction evoked by suprathreshold heat stimulation (i.e. causing heat hyperalgesia) is also able to decrease the heat threshold leading to heat allodynia. However, it is not always the case. Intraplantar injection of carrageenin evoked a swelling of the rat hindpaw and a shortening of latency time measured by a conventional, constant-temperature hot plate 1–3 hours after treatment. Surprisingly, carrageenin treatment failed to decrease the heat threshold measured by our increasing-temperature hot plate (unpublished observation) suggesting that the two parameters of thermal sensitization, the drop of heat threshold and the shortening of latency time of the response to a suprathreshold heat stimulus, are not necessarily coupled, possibly due to their different pathophysiological regulation.

Intraplantar injection of the potent TRPV1 receptor agonist RTX evoked a transient nocifensive reaction after which a considerable drop of the heat threshold measured by increasing-temperature hot plate was observed: the initial, maximal threshold drop ranged between 6–10°C followed by a nearly complete recovery by 25 min [33]. This heat allodynia was potently inhibited by i.p. applied conventional analgesics at all time points of measurement: MEDs for morphine, diclofenac and paracetamol were 1, 1 and 100 mg/kg, respectively, which are consistently lower than such values for elevating the baseline heat threshold indicating the high pharmacological sensitivity of this heat allodynia paradigm.

One of the hindpaws of superficially anesthetized rats was immersed in hot water (51°C) for 20s. This mild heat injury failed to evoke any overt nocifensive reaction after recovery from anesthesia but evoked a drop in the heat threshold for at least 4 hours as measured by the increasing-temperature water bath [124]. The drop of the heat threshold measured at different time points varied between 8 and 10°C. This heat allodynia was potently inhibited by i.p. applied conventional analgesics: MEDs for morphine, diclofenac, ibuprofen, and paracetamol were 0.3, 0.3, 10 and 30 mg/kg, respectively. In addition, local, intraplantar pretreatment with doses lower than these MEDs of morphine, diclofenac or ibuprofen were also able to inhibit the drop of the heat threshold providing evidence for peripheral sites of action of these drugs.

Under light general anesthesia, a standardized surgical incision was made on the plantar surface of the hindpaw according to ¹²⁶. After recovery from anesthesia, a 5–7°C decrease of the heat threshold lasting for at least a week became evident upon daily measurements by the increasing-temperature water bath [127]. Analgesics given i.p. after the first postoperative threshold measurement dose-dependently reduced the drop of heat threshold with similar MED values as in the case of the previous models: morphine, diclofenac, and paracetamol with 0.3, 1 and 100 mg/kg, respectively. Intraplantar pretreatment with lower doses of morphine or diclofenac was also able to evoke an antiallodynic effect by a peripheral mechanism(s).

At the time of the experiments, TRPV1 receptor antagonists were in the limelight of drug development with more than 20 compounds tested preclinically by different drug companies [128,129]. We compared 3 such compounds using the recently developed paradigms based on drop of the noxious heat threshold [130]. SB705498 [131], BCTC [132] and AMG9810 [133] were all without effect on the baseline heat threshold but able to reduce the drop of the heat threshold induced by RTX injection, mild heat injury or surgical incision. The MED values ranged between 1 and 10 mg/kg indicating a high sensitivity of these paradigms to non-conventional potential analgesics as well. Importantly, the effects of these antagonists were also assessed in the RTX model using the plantar test apparatus based on the measurement of latency time [134]. Surprisingly, in the latter, case 30 times higher MED values were found for all 3 antagonists indicating a very high sensitivity of the heat threshold measurement paradigm to the effects of TRPV1 receptor antagonists.

In diverse experimental arrangements, we have employed noxious heat threshold measurement as a reliable tool for studying mechanisms of thermonociception. These experiments aimed at examining the features of the TRPV1 receptor as well as studying peripheral mechanisms of heat allodynia evoked by different stimuli. The fatty acid amide arachidonoyl-ethanolamide (anandamide) has drawn attention because it can activate not only the cannabinoid receptors, mainly the CB₁ but TRPV1 as well [135]. In our experiments, however, anandamide applied intraplantarly failed to induce a nociceptive reaction or an alteration of the noxious heat threshold measured by the increasing-temperature water bath. Nevertheless, it could prevent heat allodynia evoked by RTX administration, and this action was abolished by a CB₁ receptor antagonist proving that the antiallodynic effect of anandamide was exclusively CB₁ receptor-mediated. Some related fatty acid amides were identified as novel TRPV1 receptor ligands partly by measurement of the noxious heat threshold [23,136]. N-oleoyl-dopamine (OLDA) injected intraplantarly evoked a transient nociceptive behavior for less than 10 min. Afterward, a 6–9°C drop of the heat threshold of the hindpaw was recorded, and this heat allodynia could be inhibited by intraplantar pretreatment with the TRPV1 antagonist iodo-resiniferatoxin indicating that OLDA is a TRPV1 receptor agonist. Similar results were obtained with 3-methyl-N-oleoyl-dopamine which can also be considered a novel TRPV1 receptor agonist. In contrast, 4-methyl-N-oleoyl-dopamine and N-oleoylethanolamide failed to evoke nocifensive behavior or drop of the heat threshold, however, neither of them prevented such responses induced by RTX suggesting a TRPV1 antagonist role for these compounds.

It is known that the phosphorylation status of the TRPV1 receptor largely determines its responsiveness to various stimuli: phosphorylation increases and dephosphorylation decreases the sensitivity of the receptor [for ref. see 137]. Protein kinase A (PKA) and C (PKC) are able to phosphorylate and thereby facilitate the function of TRPV1 [138,139]. Our aim was to assess the relative roles of these enzymes in TRPV1 sensitization using the heat threshold measuring paradigm. In our RTX-based heat allodynia model, intraplantar pretreatment with an inhibitor of PKA reduced the heat threshold-lowering effect of RTX measured at 5 min while a PKC inhibitor was without effect. In contrast, an activator of either PKA or PKC was able to increase the threshold-lowering action of RTX. The results suggest that PKA-mediated phosphorylation plays a significant role in the setting of the baseline sensitivity of TRPV1 whereas PKC does not appear to have a similar role. However, an enhancement of TRPV1 phosphorylation over the baseline level by either PKA or PKC can further enhance TRPV1 responsiveness.

Although TRPV1 can be activated by temperatures above 43°C, noxious heat responsiveness of mice lacking the gene for TRPV1 is normal, at least with low/moderate heat intensities, upon measurement of withdrawal latency of the hindpaw or tail [140,141]. Some data suggested a paradoxical involvement of the cold-sensitive TRPA1 channel in noxious heat responsiveness [142]. Therefore, we wished to assess the role of both TRPV1 and TRPA1 channels in heat responsiveness using the heat threshold measurement paradigm. The heat thresholds of the mouse hindpaw and tail were investigated in a parallel way by using the increasing-temperature hot plate and water bath, respectively [125]. Compared with wild-types, the heat threshold of the hindpaw of TRPV1 knockout mice was not different but the threshold of the tail was higher indicating a contribution of TRPV1 to the baseline noxious heat threshold of the mouse tail but not the hindpaw. The role of TRPV1 in setting the heat threshold on the tail is an important new finding because it is similar to humans in whom TRPV1 antagonists elevated the noxious heat threshold causing scald injury [128,129]. The heat threshold of the paw or tail was not different in TRPA1 knockouts compared with wild-types. This is in contrast with the results obtained by ¹⁴²,who found some involvement of TRPA1 in the heat response of the mouse paw using a different approach, i.e. latency measurement. As shown by a more recent study, the mechanism of acute noxious heat encoding involves the trio of TRPM3, TRPV1 and TRPA1 channels [143].

We have compared the mediator background of the heat allodynia associated with mild heat injury and surgical incision [144]. Test compounds (antagonists of TRPV1, bradykinin B₁, B₂ and P2X receptors and inhibitors of cyclooxygenase, lipoxygenase and nitric oxide synthase enzymes) were administered intraplantarly in order to avoid their systemic actions. By measuring the inhibition of the heat injury- or incision-evoked drop of the heat threshold, surprisingly small differences regarding the mediator background between the two models were revealed: both allodynic phenomena involve activation of the TRPV1 receptor, B₂ bradykinin receptor and P2X purinoceptor as well as formation of prostaglandins and nitric oxide. In case of the mild heat injury, additional formation of lipoxygenase products, with plantar incision additional activation of the B₁ bradykinin receptor was revealed.

After the cessation of their sensory excitatory action, certain agonists of the TRPV1 receptor (e.g. capsaicin, RTX) can induce two levels of sensory desensitization, i.e. reduced responsiveness, in a concentration- and time-dependent manner. At lower concentrations and with shorter exposures, desensitization is restricted to the TRPV1 receptor itself, i.e. the nociceptive nerve terminal shows reduced responsiveness only to stimuli acting on TRPV1. At higher concentrations and upon longer exposures, desensitization involves the entire nerve terminal, with reduced responsiveness to all stimuli including thermal, mechanical and chemical ones. Importantly, although hot (> 43°C) stimuli can activate TRPV1, this channel has no role in determining the baseline noxious heat threshold in rat or mouse hindpaw (see above). Consequently, TRPV1-specific desensitization does not change the nociceptive heat threshold, but desensitization of the whole nerve terminal is associated with an increase in the heat threshold. The aim of our studies was to compare the sensory desensitizing effects of classical (capsaicin, RTX) and newly identified (OLDA) TRPV1 receptor agonists by measuring the nociceptive heat threshold. Bilateral intraplantar treatment with capsaicin or RTX dose-dependently increased the heat threshold for more than a week measured by the increasing-temperature hot plate [145]. This is interpreted as a manifestation of desensitization of the TRPV1-expressing nerve terminal. Similar treatment with OLDA, however, failed to elevate the heat threshold. A lower dose of RTX used in our acute heat allodynia model given intraplantarly decreased the heat threshold drop evoked by a subsequent RTX injection 3 h later when the heat threshold was normal. Similar results were obtained with OLDA. These actions are regarded as a sign of TRPV1-specific desensitization.

Summary

Prof. János Szolcsányi spent 60 years studying the properties of capsaicin-sensitive sensory neurons. The existence of the latter subpopulation of primary afferent neurons itself has been adopted by the scientific community as a result of his extensive and fruitful research activity. His research group has provided substantial evidence for the triple function of the peptidergic subgroup of the capsaicin-sensitive sensory neurons including classical afferent function, local efferent tissue responses and remote, hormone-like anti-inflammatory and antinociceptive actions. This latter phenomenon was named “sensocrine” function by his own words. These discoveries have, and in the future certainly will have, broad implications for studying pathophysiological regulation of inflammation and pain as well as for the development of novel analgesic drugs acting directly on peripheral nociceptors. The immunohistological pictures shown below demonstrate co-localization of TRPV1 and TRPA1 channels, CGRP and SOM in primary sensory neurons from the L4 dorsal root ganglion of the rat using the RNAscope method (Figure 5). These four proteins/peptides are not only characteristic markers of these neurons but also cornerstones of the splendid research work performed by János Szolcsányi and his coworkers.

Figure 5.

Representative confocal images of mouse Trpa1 and Trpv1 mRNA on dorsal root ganglion (DRG). Trpa1 (green) and Trpv1 (red) mRNA co-localized with calcitonin gene-related peptide (CGRP, purple) and somatostatin (SOM, blue) mRNA on peptidergic and non-peptidergic sensory neurons of C57Bl/6 J mouse lumbar 4 (L4) DRG, counterstained with DAPI (cyan) (a). Trpa1 (green) co-localized Trpv1 (red) mRNA (b), calcitonin gene-related peptide (CGRP, purple) co-localized with somatostatin (SOM, blue) mRNA (c). Samples were imaged by using LSM 710 confocal laser scanning microscope (Carl Zeiss, Jena, Germany). Brightness/contrast adjustment was processed using Fiji, 1.53c, NIH, USA. Scale bar: 20 µm (image prepared by Angéla Kecskés).

Funding Statement

This work was funded by the Hungarian National Research, Development and Innovation Office K134214, K138046 and K138936. We acknowledge the support of the Governmental Information Technology Development Agency, Hungary. “Project no. RRF-2.3.1-21-202200015 has been implemented with the support provided by the European Union.” A.K was supported by the János Bolyai Research Scholarship of the Hungarian Academy of Sciences. Bolyai Fellowship and Research grant of Medical School, University of Pécs (KA-2021-23; Hungarian Scientific Research FundMagyarország Kormánya [RRF-2.3.1-21-2022-00015];

Abbreviations

CGRP: calcitonin gene-related peptide; DRG: dorsal root ganglion; MCD: methyl β-cyclodextrin; MED: minimum effective dose; NGF: nerve growth factor; OLDA: N-oleoyl-dopamine; PKA: protein kinase A; PKC: protein kinase C; RTX: resiniferatoxin; SMase: sphingomyelinase; SOM: somatostatin; SRIF: somatotropin release inhibiting factor; SST_1–5: somatostatin receptors; SP: substance P; TRPA1: transient receptor potential ankyrin 1; TRPM3: transient receptor potential melastatin 3; TRPM8: transient receptor potential melastatin 8; TRPV1: transient receptor potential vanilloid 1.

Disclosure statement

No potential conflict of interest was reported by the author(s).