TRPA1 mediates the antinociceptive properties of the constituent of Crocus sativus L., safranal

Abstract Safranal, contained in Crocus sativus L., exerts anti‐inflammatory and analgesic effects. However, the underlying mechanisms for such effects are poorly understood. We explored whether safranal targets the transient receptor potential ankyrin 1 (TRPA1) channel, which in nociceptors mediates pain signals. Safranal by binding to specific cysteine/lysine residues, stimulates TRPA1, but not the TRP vanilloid 1 and 4 channels (TRPV1 and TRPV4), evoking calcium responses and currents in human cells and rat and mouse dorsal root ganglion (DRG) neurons. Genetic deletion or pharmacological blockade of TRPA1 attenuated safranal‐evoked release of calcitonin gene‐related peptide (CGRP) from rat and mouse dorsal spinal cord, and acute nociception in mice. Safranal contracted rat urinary bladder isolated strips in a TRPA1‐dependent manner, behaving as a partial agonist. After exposure to safranal the ability of allyl isothiocyanate (TRPA1 agonist), but not that of capsaicin (TRPV1 agonist) or GSK1016790A (TRPV4 agonist), to evoke currents in DRG neurons, contraction of urinary bladder strips and CGRP release from spinal cord slices in rats, and acute nociception in mice underwent desensitization. As previously shown for other herbal extracts, including petasites or parthenolide, safranal might exert analgesic properties by partial agonism and selective desensitization of the TRPA1 channel.

used to treat headache. 5 Preclinical studies focused on the pharmacological activity of saffron and its purified constituents, suggesting anti-oxidant, 6 anti-inflammatory 7 and antinociceptive 8 properties of the golden spice. Antinociceptive effect of safranal has been demonstrated in models of inflammatory pain, such as formalin, acetic acid or carrageenan, 8 and in models of neuropathic pain, including chronic constriction injury and nerve crush injury. 9,10 The analgesic action of safranal has been attributed to its ability to suppress glial activation and proinflammatory cytokines production in the central nervous system. 11 However, the underlying mechanisms responsible for the analgesic action of saffron have not been yet elucidated.
Transient receptor potential (TRP) channels are pleiotropic excitatory ion channels present in a large variety of cells. Some TRPs are highly expressed by specific subpopulations of primary sensory nociceptive neurons, where they contribute to sensing noxious chemical, mechanical and thermal stimuli. [12][13][14] In particular, the TRP ankyrin 1 (TRPA1) is coexpressed with the vanilloid type 1 (TRPV1, the capsaicin receptor), or the vanilloid type 4 (TRPV4) by a subpopulation of nociceptors, which contain and release the proinflammatory neuropeptide calcitonin gene-related peptide (CGRP) and substance P (SP). 12,13 CGRP and SP released from peripheral endings of primary afferents mediate neurogenic inflammation 13 and CGRP is now recognized to play an essential role in migraine pain. 13,15 TRPA1 is activated by a number of exogenous irritant and pungent agents, including allyl isothiocyanate (AITC), cinnamaldehyde 16,17 and formalin. 18 Notably, reactive oxygen, nitrogen and carbon species gate TRPA1 by reacting with specific cysteine/lysine residues, thus mediating pain and neurogenic inflammation. [19][20][21][22][23] Here, we evaluated whether the three main constituents of saffron affect the function of TRP channels expressed in nociceptors, and in particular that of TRPA1. We found that safranal and its precursor, picrocrocin, but not crocin, selectively activate TRPA1, but not TRPV1 or TRPV4, safranal being far more effective than picrocrocin. In addition, safranal behaves as a partial agonist on TRPA1 as compared to the full agonist, AITC. Finally, safranal induces selective desensitization of the TRPA1 channel, thus attenuating neuronal excitation that results in nociception and CGRP release. This unforeseen TRPA1-desensitization mechanism could possibly explain the analgesic effects attributed to saffron.

| Animals
In vivo experiments and tissue collection were carried out according to the European Union (EU) directive guidelines (2010/63/EU) and Italian legislation (DLgs 26/2014) for animal care procedures. Studies were conducted under the University of Florence research permit #194/2015-PR. Animals were housed in a temperature-and humidity-controlled vivarium (12-hour dark/light cycle, free access to food and water) and were allowed to acclimatize for at least 7 days prior to experimentation. Animal studies were reported in compliance with the ARRIVE guidelines. 24,25 A group size of six animals for behavioural experiments was determined by sample size estimation using G*Power (v3.1) 26 to detect size effect in a post-hoc test with type 1 and 2 error rates of 5 and 20% respectively. Allocation concealment was performed using a randomization procedure (http://www.rand omizer.org/).

| Organ bath assay
Rat urinary bladder was excised from rat, and longitudinal strips were suspended at a resting tension of 1 g in 10-mL organ bath

| Drugs and reagents
acetamide] was synthesized as previously described. 32 If not otherwise indicated, all other reagents were obtained from Sigma-Aldrich (Milan, Italy).

| Statistical analysis
Statistical analysis was performed using the unpaired two-tailed Student's t test for comparisons between two groups and the one-or two-way ANOVA, followed by the post-hoc Bonferroni's test for comparisons of multiple groups (GraphPad Prism version 5.00, San Diego, CA, USA). A P < 0.05 was considered statistically significant.
Data are reported as mean ± SEM.

| Safranal and picrocrocin, but not crocin, selectively activate the human TRPA1 channel
Safranal and picrocrocin evoked a concentration-dependent calcium response in hTRPA1-HEK293 (EC 50s 17 ± 0.3 μmol/L and 56 ± 0.3 μmol/L respectively), but not in naïve HEK293 cells ( Figure 1B). The selective TRPA1 antagonist HC-030031 attenuated the response evoked by safranal and picrocrocin and AITC ( Figure 1B). Crocin, the third main constituent of saffron was not investigated in this test because of its intense yellow colour which interferes with the recording system. Thus, the ability of crocin to target TRPA1 was analysed by whole-cell patch-clamp electrophysiology. Crocin did not evoke any measurable inward current in hTRPA1-HEK293, compared to its vehicle ( Figure 1C).
Further investigation was limited to safranal, the most potent of the three compounds. The non-electrophilic agonist, menthol, evoked a robust calcium response in 3C/K-Q hTRPA1-HEK293 cells, which express a mutant form of the TRPA1 receptor, lacking three key cysteine (C619, C639, C663) and one lysine (K708) residues, which are required for channel activation by electrophilic agonists 23,33 ( Figure 1D).

| Safranal selectively excites TRPA1 in rodent sensory neurons
Exposure to safranal of cultured rat DRG neurons evoked a concentration-dependent calcium response in a subset of cells, identified as nociceptors by their ability to respond to KCl, capsaicin and AITC 31 ( Figure 2A). In rat DRG neurons maximum calcium responses to safranal or AITC were 35.8% ± 8.5% and 56.0% ± 6.0% of ionomycin (n = 25, P < 0.01) respectively, and EC 50s were 38 ± 0.03 μmol/L and 5 ± 0.3 μmol/L respectively. Calcium responses were attenuated by HC-030031, and unaffected by the respective TRPV1 and TRPV4 selective antagonists, capsazepine and HC-067047 ( Figure 2A).
Safranal or AITC elicited inward currents in rat DRG neurons, that were blocked by HC-030031 ( Figure 2B), which, however, did not affect currents evoked by capsaicin, indicating selectivity (Figure 2A,   B). Safranal and AITC produced calcium responses in DRG neurons isolated from Trpa1 +/+ mice, but not in neurons isolated from Trpa1 −/− mice ( Figure 2C), while the response to capsaicin was unchanged in both mice strains.

| Safranal causes neuropeptides release and acute pain via TRPA1 activation in nociceptors
There is evidence that TRPA1 is localized in peptidergic primary sensory neurons. 35  and SP. SP released upon stimulation of TRPV1 or TRPA1 results in a contractile response of isolated strips of rat urinary bladder that is mediated by activation of the NK1 and NK2 receptors for SP in bladder smooth muscle cells. 37 Safranal caused a concentration-dependent contractile response of isolated strips of rat urinary bladder with a slightly lower potency than AITC (EC 50s were 76 ± 0.07 μmol/L and 56 ± 0.02 μmol/L respectively) ( Figure 3A). Efficacy of safranal was also lower than F I G U R E 1 Safranal (SFR) and picrocrocin (PICR) selectively target the human TRPA1 channel. A, Chemical structures of SFR, PICR and crocin (CRO). B, Concentration response curves of the calcium mobilization evoked by SFR and PICR in hTRPA1 HEK293 cells. Representative traces and pooled data of calcium response evoked by SFR, PICR and AITC in hTRPA1 HEK293 pre-exposed to HC-030031 (HC03, 30 μmol/L) or its vehicle (-) and in naïve HEK293 cells. C, Representative traces and pooled data of whole-cell patch-clamp inward currents evoked by CRO and AITC (100 μmol/L) in hTRPA1 HEK293. D, Pooled data of calcium responses evoked by SFR and menthol in wild-type (wt) and mutant (3C/K-Q) hTRPA1 HEK293 transfected cells. E, Concentration response curves of the calcium mobilization evoked by SFR and PICR in IMR90 cells. Representative traces and pooled data of calcium responses evoked by SFR, PICR and AITC pre-exposed HC03 (30 μmol/L) or its vehicle (-) in IMR90 cells. HC03 does not affect the response evoked by the hPAR 2 -AP (100 μmol/L). Veh is the vehicle of SFR, dash (-) indicates the vehicle of HC03. Data are mean ± SEM of n > 20 cells from 4 to 6 independent experiments (B, D, E) and n > 3 cells from 3 to 5 independent experiments (C). *P < 0.05 vs veh, § P < 0.05 vs SFR, PICR or AITC; one-way ANOVA with Bonferroni post-hoc correction that of AITC (maximum response, 15.5% ± 3.0% of carbachol, n = 6, and 35.1% ± 6.2% of carbachol respectively, n = 6, P < 0.05) (Figure 3A). The response of safranal was attenuated by pre-exposure to a combination of NK1 and NK2 receptor antagonists (L-733,060 and SR48968 respectively), by pre-exposure to a high concentration of capsaicin able to desensitize the nociceptors and by pre-exposure to the TRPA1 selective antagonist, HC-030031, while it was unaffected by pre-exposure to capsazepine (TRPV1 antagonist) and HC-067047 (TRPV4 antagonist) ( Figure 3A). This finding supports the hypothesis that safranal behaves as a partial agonist at the TRPA1 channel.
Safranal elicited a concentration-dependent increase in CGRP-LI outflow from rat spinal cord slices which was absent in a calciumfree medium or in tissues pre-exposed to a desensitizing concentration of capsaicin and in the presence of HC-030031 ( Figure 3B).

| In vitro and in vivo exposure to safranal attenuated TRPA1-mediated responses
In hTRPA1-HEK293 transfected cells, inward currents evoked by AITC underwent a concentration-dependent attenuation after preexposure to increasing concentrations of safranal ( Figure 4A), whereas KCl-evoked currents were not affected ( Figure 4A). In cultured rat DRG neurons, pre-exposure to safranal reduced inward currents evoked by AITC, but not currents evoked by capsaicin, suggesting that safranal promotes selective TRPA1 desensitization (Figure 4B). Specific desensitization of TRPA1 was also observed in organ bath experiments. Pre-exposure to safranal reduced contractile responses of rat urinary bladder strips evoked by safranal and AITC, but not those evoked by capsaicin, GSK1016790A or carbachol (Figure 4C). Finally, pre-exposure to safranal was able to markedly reduce CGRP release from rat spinal cord evoked by AITC, without affecting the release evoked by capsaicin or GSK1016790A ( Figure 4D).
To test whether pre-exposure to safranal could desensitize TRPA1-mediated responses in vivo, safranal was administered by the intragastric route (i.g.) at two different doses (0.5-1 mg/kg). One sin- indicating that, similarly to other electrophilic and reactive agonists, 23,33 channel activation by the saffron constituent is mediated by specific cysteine/lysine residues. Both potency and efficacy of the calcium response evoked by safranal in cultured rat DRG neurons were lower than that of AITC. In particular, the lower efficacy supports the hypothesis that safranal may act as a partial TRPA1 agonist.
Peptidergic sensory neurons expressing TRPA1 exert a dual afferent and efferent role, because upon stimulation they can both signal pain to the brain and release from their peripheral terminals the neuropeptides SP and CGRP, which mediate neurogenic inflammatory responses. 13,39 One of these responses is the SP-mediated contraction of rat urinary bladder smooth muscle, 40 which can be also produced by TRPA1 activation in intramural sensory nerve terminals. 37 Here, we found that safranal produced a contraction mechanistically similar to that induced by AITC, because it was abrogated after exposure to a high, desensitizing, concentration of capsaicin, in the presence of a TRPA1 selective antagonist and in the presence of a combination of NK1 and NK2 receptor antagonists. Notably, the efficacy of safranal to target TRPA1 in the urinary bladder was lower than that of AITC, further supporting a possible partial agonism of the compound.
Results obtained in vitro were recapitulated in an in vivo setting, where the proalgesic action of safranal was mediated exclusively by TRPA1. Intraplantar injection of safranal elicited a concentration-dependent acute nociceptive response that was TRPA1-dependent, being selectively attenuated by TRPA1 antagonism and gene deletion. However, there is also indication that saffron has beneficial effect in certain pain conditions. 3,41 Furthermore, safranal has been reported to attenuate pain-like responses in animal models of inflammatory and neuropathic pain. [8][9][10] In particular, safranal attenuated mechanical allodynia and thermal hyperalgesia in a chronic constriction injury model, 10 19 Thus, it is possible that safranal exerts its partial agonist and desensitizing activities at the Schwann cell TRPA1, contributing in this manner to reduce neuropathic pain.
Clinical investigations with saffron has reported headache as a possible adverse reaction. 4 In contrast with this observation, saffron has been used by Indian traditional medicine to treat headaches. 5 TRPA1 is preferentially expressed by peptidergic sensory neurons, and upon its activation evokes the simultaneous release of the proinflammatory and proalgesic neuropeptides, SP and CGRP. 21,43 CGRP is now considered a major contributor of migraine pain as small molecule CGRP receptor antagonists and monoclonal antibodies against CGRP or its receptor have marked beneficial effects in migraine. 13,15 The observation that in strips rat urinary bladder and in rat dorsal spinal cord slices pre-exposure to safranal reduced AITC-evoked SP-mediated contractile responses and CGRP release respectively, without affecting TRPA1-independent responses further suggests that safranal attenuates the efferent and pro-migraine function of peptidergic nociceptors.
We previously showed that upon in vitro or in vivo exposure to other herbal preparations, such as isopetasin, contained in