Characterization of the anandamide induced depolarization of guinea-pig isolated vagus nerve

Introduction

The existence of a capsaicin receptor was suggested almost 30 years ago (Szolcsányi & Jancsó-gábor, 1975) and with the discovery of resiniferatoxin as an ultrapotent capsaicin analog and capsazepine as a vanilloid receptor (VR) antagonist, considerable advances have been made concerning the role of VR (Szallasi & Blumberg, 1990; Bevan et al., 1992). Recently, VR1 has been cloned (Caterina et al., 1997) and targeted disruption of the VR1 gene in mice resulted in an impairment in response to high doses of capsaicin and is implicated in hyperalgesia (Caterina et al., 2000; Davis et al., 2000).

VR1 is a non-selective cation channel with a high calcium permeability that is activated by a broad spectrum of stimuli, including capsaicin, heat, low pH and the endogenous cannabinoid anandamide (Caterina et al., 1997; Tominaga et al., 1998; Zygmunt et al., 1999). We and others have recently shown that the endogenous cannabinoid (CB) receptor agonist anandamide produces a concentration-dependent contraction of guinea-pig isolated bronchus via the activation of the VR1 receptor in this preparation (Craib et al., 2001; Tucker et al., 2001). However, the mechanism by which anandamide affects airways is far from resolved. The vasodilator response to anandamide in isolated arteries is capsaicin-sensitive and accompanied by release of calcitonin-gene-related peptide (CGRP) (Zygmunt et al., 1999). Furthermore, anandamide has been reported to inhibit bronchospasm and cough evoked by capsaicin in the guinea-pig, but to elicit bronchospasm when the constricting tone exerted by the vagus nerve is removed, both effects being mediated through activation of peripheral CB1 cannabinoid receptors (Calignano et al., 2000). In contrast, it has recently been suggested that products derived from the metabolism of anandamide by lipoxygenase leads to the activation of the vanilloid receptor in airway tissue (Craib et al., 2001).

In the present study we have further investigated the role of the vanilloid receptor in the responses to anandamide in isolated vagal airway preparations. We have also investigated the role of cannabinoid receptors, protein kinase C and the lipoxygenase pathway on the response to anandamide in this vagal preparation.

Methods

Results

Discussion

In the present study, we have provided further evidence that the putative endogenous cannabinoid agonist anandamide can activate vanilloid receptors. Our results show that anandamide is able to induce depolarization of guinea-pig isolated vagus nerve preparations, which although less than that induced by capsaicin, was inhibited by the VR1 antagonist capsazepine, but not by the CB1 receptor antagonist SR141716A or the CB2 receptor antagonist SR144528 and may in part be due to the generation of lipoxygenase products. Consistent with the observation that the depolarization of isolated vagus nerve was CB1/CB2 receptor independent, the cannabinoid receptor agonist, CP55940 failed to depolarize the vagus nerve. Our work also confirms that the anandamide-induced response was attenuated by capsaicin pretreatment, while anandamide failed to cross-desensitize the response to capsaicin. Moreover the ability of the PKC activator PMA to augment responses to both capsaicin and anandamide also points to a common mechanism of action.

The grease gap recording technique has previously been used to measure depolarization of vagus nerve to various substances (Ireland & Tyers, 1987; Rang & Ritchie, 1988; Butler et al., 1990). Using a similar technique, we have demonstrated reproducible depolarization responses to KCl, capsaicin, anandamide and arachidonic acid. This technique offered us the advantage of studying the activation of VR1 receptors directly, rather than indirectly via the release of neuropeptides and subsequent contraction of bronchial smooth muscle (Tucker et al., 2001), although this technique does not allow one to distinguish between the type of afferent nerves being activated. Recently, it has been shown that anandamide activates cardiac vagal nerve endings (Malinowska et al., 2001) and pulmonary vagal C-fibre afferents (Lin & Lee, 2002) in anaesthetized rats via a vanilloid receptor dependent mechanism. In contrast, anandamide did not activate rapidly adapting pulmonary receptors (RARs), suggesting that anandamide selectively excites C-fibres (Lin & Lee, 2002). However in the guinea-pig, capsaicin can induce depolarization of both vagal C and Aδ fibres whose cell bodies reside in the jugular ganglion (Holzer, 1991; Szallasi, 1995a; Riccio et al., 1996; Kajekar et al., 1999). Therefore, in the present study, our confocal microscopy observations have provided clear evidence that VR1 receptors are present on both neuropeptide containing nerves as well as nerves that do not stain for substance P. The nature of the neurotransmitter in the VR1 positive nerves that do not stain substance P remains to be established, but it is likely that both would be stimulated by anandamide. In 1997, the vanilloid receptor was identified by using an expression cloning strategy based on calcium influx to isolate a functional cDNA encoding a capsaicin receptor from sensory neurons (Caterina et al., 1997). VRs are expressed along the entire length of vanilloid-sensitive sensory neurons (Szallasi, 1995a) and nerve ligation studies suggest that VRs are transported from the cell bodies to the periphery in a form capable of ligand binding (Szallasi et al., 1995b) and it is also clear that these receptor are functionally expressed along vagal nerve fibres as demonstrated in this study.

Both cloned (VR1) and native VRs are non-selective cation channels that are gated by heat, low pH and vanilloid agonists (Tominaga et al., 1998). However, in recent years considerable attention has been focused on identifying endogenous ligands for the vanilloid receptor. It has recently been demonstrated that certain lipoxygenase products of arachidonic acid metabolism, in particular 12- and 15-(S)-hydroperoxyeicosatetraenoic acids, 5- and 15-(S)-hydroxyeicosatetraenoic acids and leukotriene B4 can activate the VR1 receptor (Hwang et al., 2000). Furthermore, anandamide displays a high structural similarity to the vanilloids, especially olvanil, a nonpungent, capsaicin analogue (di marzo et al., 1998). It is perhaps not surprising therefore, that the putative endogenous cannabinoid receptor agonist anandamide is also capable of activating VR1 receptors in transfected cells (Zygmunt et al., 1999; Smart et al., 2000; Ross et al., 2001), DRG neurones (Hwang et al., 2000; Smart et al., 2000; Tognetto et al., 2001) and in airway tissue (Craib et al., 2001; de petrocellis et al., 2001b; Tucker et al., 2001). Recently, these findings have been extended with the suggestion that anandamide itself may be metabolized by a lipoxygenase-dependent pathway, and whose products activate the vanilloid receptor (Craib et al., 2001). A number of biochemical studies have shown that hydroxyl derivatives of anandamide are formed by 12-,15- but not 5-lipoxygenase (Ueda et al., 1995; Hampson et al., 1995; Edgemond et al., 1998). The pharmacology of these derivatives has been investigated and like anandamide, they bind to CB1-receptors (Edgemond et al., 1998). Interestingly, the 15-hydroxy derivative of anandamide is considerably less potent than anandamide in stimulating the vanilloid receptor. Therefore, the role of these metabolites as endogenous activators of the vanilloid receptor is unclear (De Petrocellis et al., 2001a), although it remains to be established whether the 12-hydroxy derivative activates the vanilloid receptor.

Anandamide is metabolised by fatty acid amide hydrolase (FAAH) to arachidonic acid (Cravatt et al., 1996) and lipoxygenase products generated from this pathway could have the potential to activate the vanilloid receptor (Hwang et al., 2000; Olah et al., 2001). Thus, we have demonstrated that anandamide induces depolarization of vagal nerves, which in part, is mediated by lipoxygenase metabolism. However, the contribution of lipid mediators derived from the lipoxygenase metabolism of anandamide (Craib et al., 2001) or of arachidonic acid, following the metabolism of anandamide by FAAH, in this response remains to be established.

In order to further explore the mechanism by which anandamide stimulates the vanilloid receptor, experiments were performed in the absence of calcium. Under such conditions, anandamide would still be subject to metabolism by FAAH (Hillard et al., 1995; Karava et al., 2001). Whilst depolarization to anandamide was unaffected by calcium removal, lipoxygenase inhibition failed to inhibit this response suggesting that anandamide directly activated native vanilloid receptor under these conditions. A further possibility is that calcium entry via the vanilloid receptor might also facilitate the mobilization of arachidonic acid with subsequent metabolism by lipoxygenase, resulting also in the activation of the vanilloid receptor. This possibility was supported by the fact that, although anandamide was previously shown to cause arachidonate release via CB1- and CB2-receptors (Shivachar et al., 1996; Hunter & Burstein, 1997; Pestonjamasp & Burstein, 1998), we found that depolarization of vagal nerves by anandamide was inhibited by the VR1 antagonist capsazepine, but not by CB1- or CB2-selective antagonists. Furthermore, the possibility that VR1-mediated calcium entry leads to arachidonate mobilization is supported also by previous findings in several cell lines (Maccarrone et al., 2000). Finally, in support of this hypothesis, we found that when the calcium requirement of arachidonate release was bypassed by treating our preparations directly with arachidonic acid, the subsequent depolarization of vagal nerves was sensitive to the lipoxygenase inhibitor ETI. However, we also observed that depolarization of vagal nerves by capsaicin which is a more potent agonist at native VR1 than anandamide, was unaffected by ETI, although in a previous study it was observed that capsaicin-induced contraction of guinea-pig bronchial preparations tended to be suppressed by lipoxygenase inhibition (Craib et al., 2001), although this was not significant. Clearly, further work is required to determine the importance of calcium entry following vanilloid receptor activation, on the generation of endogenous lipid activators of this ion channel.

There is currently some dispute as to the physiological significance of the role of anandamide as a CB and VR selective agonist, as the high concentration of anandamide reported to be required in order to activate the VR1 receptor is greater than that required to activate CB receptors (Szolcsányi, 2000). Whether local concentrations generated within the microenvironment of the nerve are sufficient to activate VR receptor is unknown at present, although it remains plausible that sensitization of afferent nerves may augment the response to endogenously liberated anandamide (di marzo et al., 2001; Olah et al., 2001; Vellani et al., 2001).

Alternatively, the high concentrations required to activate native vanilloid receptor might be a consequence of the metabolism of anandamide by lipoxygenase (Craib et al., 2001), by FAAH to arachidonic acid and subsequent metabolism by lipoxygenase and/or the generation of lipoxygenase products via a calcium dependent process. Further studies measuring the release of lipoxygenase products generated by these different pathways in sensory nerves will be required for definitive proof of this mechanism.

We have previously reported that the contractile response to anandamide in guinea-pig isolated bronchus was inhibited by the NK2-selective antagonist, SR48968 but not by the NK1-selective antagonist, SR140333 (Tucker et al., 2001). These results provided indirect evidence that anandamide induced contraction of airway smooth muscle was a consequence of activating VR1 receptors on peripheral sensory nerves with subsequent release of endogenous neuropeptides. Here we provide further direct evidence that anandamide can activate VR1 receptors by showing that depolarization to anandamide was inhibited by the VR1 receptor antagonist capsazepine. Similarly, the vanilloid receptor antagonist capsazepine inhibited anandamide-induced vasodilation, release of CGRP and activation of the cloned receptor (Zygmunt et al., 1999). Whilst capsazepine is routinely used as a vanilloid receptor antagonist there is evidence that this agent may also induce a time dependent block of voltage-gated calcium channels (Docherty et al., 1997) and nicotinic acetylcholine (nACh) receptors (Liu & Simon, 1997). Potassium (Kuenzi & Dale, 1996) and hyperpolarization-activated cyclic nucleotide gated (HCN) currents (Gill et al., 2002) are also inhibited by capsazepine. It would seem unlikely that nACh receptors or HCN channels are present in the axons in the vagus nerve and contribute to the capsaicin-induced depolarization. On the other hand, if the capsaicin-induced depolarization is dependent on subsequent activation of voltage-gated sodium channels then this could confound interpretation of the data. However, capsaicin-mediated depolarization in mammalian vagus is still present in sodium free solution (Marsh et al., 1987). Furthermore, we have demonstrated that depolarization to capsaicin is not blocked by the removal of extracellular calcium. Therefore, voltage-gated sodium and calcium channels do not contribute greatly to the responses observed to capsaicin.

It is known that capsaicin causes a rapid desensitization of vanilloid receptors (Szallasi & Blumberg, 1999). In the present study, pretreatment with capsaicin to cause desensitization of VR1 receptors also abolished the anandamide-induced depolarization response as well as the capsaicin-induced depolarization response in the isolated vagal nerve preparation. However, anandamide failed to cross-desensitize the depolarization response to capsaicin, confirming our earlier results with guinea-pig isolated bronchus (Tucker et al., 2001). It is of interest that 64 and 72% of capsaicin sensitive afferent fibres was activated by anandamide in normal and arthritic rat joints respectively (Gauldie et al., 2001), which is similar to our findings that only 61% of capsaicin sensitive nerves responded to anandamide. This suggests that anandamide is less efficacious than capsaicin in stimulating the vanilloid receptor as the inward current of anandamide was significantly smaller (<50%) than that of capsaicin (Zygmunt et al., 1999; Hwang et al., 2000; Smart et al., 2000). Similarly, at normal pH anandamide behaves as a partial agonist (Olah et al., 2001) and therefore would be relatively poor at inducing desensitization to an agonist with greater efficacy like capsaicin. In relation to desensitization of the capsaicin response to repeated stimulation with this agonist, it was reported that acute desensitization was facilitated by calcium entry (Docherty et al., 1996; Koplas et al., 1997), a finding which we have confirmed in guinea-pig isolated vagus nerve.

There is a possibility that gating and phosphorylation of VR1 is regulated by Ca²⁺, neurotrophin receptor tyrosine kinase, protein kinase A and protein kinase C (Kress & Guenther, 1999; Winston et al., 2001; Lopshire & Nicol, 1998; Vellani et al., 2001). Nociceptors are either coupled directly to ionotropic P2X receptors or to the G protein-mediated metabotropic P2Y receptors and most P2Y receptors are coupled to Ca²⁺-sensitive (PKC-β, γ) or insensitive (PKC-ε, δ) forms of PKC (Premkumar, 2001). Bradykinin sensitizes the heat-gated current by activating PKC-ε (Cesare et al., 1999), and PKC-ε gene has a role in regulating thermal and acid-induced hyperalgesia (Khasar et al., 1999). On the other hand, PKC-γ also appears to play a role in regulating neuropathic pain (Malmberg et al., 1997), and extracellular ATP potentiates VR1 responses through metabotropic ATP (P2Y1) receptors in a PKC-dependent manner (Tominaga et al., 2001). These results suggest that PKC may link a range of stimuli to the activation of VR (di marzo et al., 2001). Similarly, activation of the vanilloid receptor by anandamide was PKC-dependent in rat VR1 expressing Xenopus laevis oocyte and neonatal rat dorsal root ganglia cell (Premkumar & Ahern, 2000). Furthermore, phosphorylation of VR1 by PKC increases the probability of channel gating by agonists including anandamide in neonatal rat dorsal root ganglia cells and in HEK293 cells expressing human vanilloid receptors (Vellani et al., 2001). In the present study, we have shown that PKC plays a role in augmenting the vanilloid response to anandamide using an electrophysiological technique in guinea-pig vagus nerve, consistent with recent findings in guinea-pig isolated bronchus (de petrocellis et al., 2001b).

In conclusion, we have demonstrated that the endogenous cannabinoid receptor agonist, anandamide evokes depolarization following activation of vanilloid receptors in guinea-pig vagus nerve that in part may also be due to the liberation of lipoxygenase products. The finding of vanilloid receptors on neuropeptide and non-neuropeptide nerves suggests that endogenous activation of this receptor generated in the lung may activate these different nerve types.