**3. Discussion**

The pathways of trigeminal neuropathic pain are poorly understood. Experimental evidences suggest a strong involvement of TRPA1 in different patterns of neuropathic pain, and recently its role was also demonstrated in a trigeminal neuropathic pain model [17].

Here we evaluated the role of TRPA1 channels in an animal model of trigeminal neuropathic pain (IoN-CCI model), investigating the effects of the TRPA1 antagonist ADM\_12 on mechanical allodynia, and neurochemical and transcriptional changes.

ADM\_12 was previously shown to revert *in vivo* the Oxaliplatin-induced neuropathy [22]. At the trigeminal level, ADM\_12 was able to reduce orofacial pain in a model of temporomandibular joint inflammation [23], and to counteract trigeminal hyperalgesia in a model of migraine pain, together with decreased *Trpa1* and neuropeptide mRNA expression levels in specific areas implicated in trigeminal pain [24].

#### *3.1. Behavioral Response*

Infraorbital nerve injury in rats leads to the development, in the ipsilateral side, of a hyporesponsiveness to mechanical stimulation within the first week post operation, followed by a hyperresponsiveness, that according to several studies [18,25], reflects a condition of mechanical allodynia. This biphasic response is probably related to the demyelination process, occurring in the early post-operative period, and remyelination process, that occurs in the late post-operative period [26]. Compared to the above cited papers [18,25,26], the time needed in this study to develop allodynia was somewhat longer. This may have been the result of small differences in the degree of nerve constriction; indeed, different degrees of IoN constriction have been shown to produce different time courses in isolated face grooming behavior [27], and this can also be true for mechanical allodynia.

The allodynic response of operated rats was abolished after treatment with the TRPA1 antagonist ADM\_12, suggesting that the blockade of TRPA1 channels located on the trigeminal afferents prevented the release of neuropeptides (CGRP and SP) [28,29], thus resulting in a reduced neurogenic inflammation, and ultimately nociceptor sensitization [30]. Accordingly, Wu and colleagues reported an increase in TRPA1 protein, as well as TRPV1 channels, in the Sp5C region of rats that underwent IoN-CCI surgery [31], confirming their involvement in this process. An additional mechanism is represented by the reduction in the release of pro-inflammatory factors via the inhibition of TRPA1 located on glial cells in the nervous system, as suggested by our results, in which we observed a reduction of the *IL-1beta*, *IL-6*, and *TNF-alpha* transcripts that possibly parallel protein expression [32–34], which could account for reduced glial cells activation [32]; or via the inhibition of TRPA1 located on non-neuronal cells, such as keratinocytes and macrophages, in the tissues surrounding the damaged nerve [35]. Pro-inflammatory mediators released in the tissues that surround the damaged nerve, and glial cell activation, are indeed known to play a crucial role in the pathophysiology of neuropathic pain [36,37]. Glial activation and pro-inflammatory cytokines are associated with the onset of neuropathic pain symptoms such as allodynia or hyperalgesia [38–42].

The involvement of TRPA1 in mechanosensation has been extensively studied; both genetic deletion of TRPA1 and its pharmacological blockade abrogate mechanical pain-like behaviors [17,43,44]. Recently, Trevisan and colleagues [17] confirmed the critical role played by TRPA1 channels in mechanical allodynia induced by trigeminal neuropathic pain; conversely, in a model of sciatic nerve injury, Lehto and co-workers [45] reported a non-significant involvement of these channels in the mechanical sensitivity. On the other hand, other authors showed that TRPA1 blockade attenuated mechanical hypersensitivity following spinal injury [46,47] or neuropathic pain induced by chemotherapeutic agents [48,49]. Altogether these observations suggest that mechanical allodynia might be differently mediated by TRPA1 channels depending on the type of pain, site of damage, or distribution profile in TG and DRGs [50]. Moreover, the different responses observed in the experimental models could also be related to the different TRPA1 antagonists used, that may inhibit the channel through binding at different sites, with specific regulatory mechanisms [51].

## *3.2. Trpa1 and Trpv1 mRNA Expression*

Chronic constriction injury of the IoN produced a marked increase in the *Trpa1* and *Trpv1* mRNA expression in central and peripheral areas ipsilaterally, and a slight increase even at the contralateral side, compared to the sham group. This contralateral increase is probably due to activation of inflammatory processes occurring after nerve injury, which can also affect the contralateral side [52]. The elevated TRP transcripts are accompanied by increased *IL-1beta*, *IL-6*, and *TNF-alpha* mRNA levels in the medulla region, and ipsilateral TG and CSC. It is known that TRPA1 and TRPV1 channels can be sensitized by inflammatory agents, causing up-regulation of these channels [53–55]. For example, *Trpa1* expression has been shown to be up-regulated by TNF-alpha and IL-1 alpha via transcriptional factor hypoxia-inducible factor-1α [56]. Similarly, TNF-alpha can up-regulate TRPV1 protein and mRNA in DRG and TG neurons [57,58]; one of the suggested pathways for *Trpv1* regulation is the p38 mitogen-activated protein kinase pathway [59], which may also be partly involved in *Trpa1*

expression [60]. As regards TRPA1, its activation seems to depend on the activation of the nuclear factor-κB signaling pathway [61].

Furthermore, an important role in neuropathic pain seems to be played by oxidative stress [62–64], whose components can directly activate TRPA1 channels [65], thereby contributing to inflammation in a TRPA1-dependent manner. Indeed, it was recently found that trigeminal neuropathic pain behaviors were mediated by TRPA1 targeted by oxidative stress by-products released from monocytes and macrophages surrounding the site of the nerve injury [17].

In agreement with our study, an up-regulation of *Trpa1* and *Trpv1* mRNA levels, as well as protein levels in TG, DRGs, and dorsal horns, has been seen in different models of neuropathic pain [11,47,60,66–71]. Increased mRNA levels may reflect an increase in functional TRPA1 and TRPV1 channels [72,73].

The increased mRNA levels detected in our experiments in CSC and medulla may have different origins: *Trpv1* mRNA undergoes bidirectional axon transport along primary afferents [74], and the same could be true for *Trpa1*, since both TRPV1 and TRPA1 are (co-)expressed, not only on peripheral, but also on central terminals of primary afferent neurons where their activation can lead to the release of transmitters that promote the sensitization of postsynaptic pain transmission pathways [75–78]. In addition, *Trpv1* mRNA could originate from GABAergic interneurons and glial cells in the rat dorsal horn, which are known to express TRPV1 [69,79].

Systemic administration of ADM\_12 markedly reduced the mRNA expression levels of both TRPs induced by IoN ligation. The effect of drug treatment on mRNA transcripts is likely to be due to an indirect effect rather than a direct one. It can be reasonably hypothesized that the effect of ADM\_12 on TRPA1 mRNA expression is indirectly due to the blockade of the channel, located either on neuronal and non-neuronal cells, which is followed by two events. On one side, the reduction of calcium (Ca2+) entry provokes a reduced activation of second messenger (Ca2+ dependent) molecules (e.g., via the phospholipase C/Ca2+ signaling pathway and Ca(2+)/calmodulin-dependent protein kinase II [CaMKII]) and interfering with the Ca2+-interacting proteins [80,81], with the consequent reduction in transcriptional rate; for example, through the CaMK—cAMP response element-binding protein (CaMK—CREB) cascade. The other event that follows TRPA1 antagonism is the reduction in neuropeptide (CGRP and SP) release [28,29], and pro-inflammatory agents from neuronal fibers and non-neuronal cells. In this frame, we hypothesize that ADM\_12 may break off a self-feeding loop in which TRPA1 channels are directly activated or sensitized by Ca2+ [51,81], endogenous substances produced by intracellular Ca2+ elevation [82], and pro-inflammatory molecules [83–85], and indirectly by the activation of nociceptive fibers caused by neuropeptide-induced neuroinflammation.

Moreover, we can also speculate that since TRPA1 and TRPV1 functions may be influenced by each other [20,86,87], a re-organization in the expression and nature of these channels after nerve injury [88–90] enabled ADM\_12 to modulate TRPV1 channels as well. Although a physical interaction between these two channels may be questionable, even if some studies described it *in vitro* [19,21], many studies reported a functional interaction between them [20,86,91–93]. For instance, Masuoka et al. [87] showed in DRG neurons that TRPA1 channels suppress TRPV1 channel activity, possibly through the regulation of basal intracellular calcium concentration, and that the TRPA1 sensitization, induced by inflammatory agents, enhance TRPV1-mediated currents [87].

These observations, including our data, show a relationship between these two TRP channels, although more information and studies are needed to understand the precise mechanisms of this putative interaction.

#### *3.3. Neuropeptide Expression*

After nerve injury, an inflammatory process leads to the release of many pro-inflammatory mediators, which participate in peripheral sensitization, promoting an excessive release of neurotransmitters [94]. Together with the inflammatory process, neuropeptides and degenerative changes affecting the nervous fibers are also crucial peripheral mechanisms [95].

In our experimental setting, the mRNA expression levels of genes coding for CGRP (*Calca*) and SP (*PTT-A*) markedly increased in the central areas containing the Sp5C, as well as in the TG ipsilateral to the IoN ligation. Interestingly, *Calca* mRNA expression in IoN ligated rats was also elevated on the contralateral TG. It has been shown that projections from the TG reach the medullary and cervical dorsal horns on both sides [96,97], and that unilateral TG stimulation activates neurons in both ipsiand contralateral Sp5C [98,99].

One of the mechanisms that could contribute to neuropeptide expression is the CaMK—CREB cascade, which is probably triggered following TRP channel activation [100], and that may represent the target mechanism for the observed inhibitory effect of ADM\_12 on the mRNA expression of CGRP and SP. The blockade of TRP channels, which co-localize with CGRP and SP in the trigeminal neurons [7,101], can inhibit *Calca* and *PPT-A* mRNA expression, thus reducing the neuropeptide release and the trigeminal sensitization process. The data supports the pivotal involvement of CGRP and SP in the delivery and transmission of pain sensation to the central nervous system, and their role in trigeminal pain syndrome. In fact, an increased concentration of neuropeptides was found in the cerebrospinal fluid and venous blood of patients with trigeminal neuralgia compared to healthy controls [102,103].

In this frame, it was quite surprising that we did not detect any significant difference in neuropeptide protein expression at the Sp5C level, neither among groups, nor between sides. Lynds and co-workers [104] reported no differences in neuropeptide (CGRP and SP) levels between ipsi- and contralateral TG two weeks after IoN transection injury, while Xu and colleagues [105] described a reduction of CGRP and SP protein levels in the ipsilateral caudal medulla eight days after partial IoN ligation. Taken together, these findings suggest that in our model the neuropeptide release at central sites might have taken place at early time points after surgery, and therefore went undetected since we only measured it on day +27, or alternatively, that CGRP and SP are mostly involved at the peripheral terminals [26]. These apparently contrasting findings prompt the need for specifically targeted studies in order to investigate in more depth the role of neuropeptide release in central and peripheral sites in this model of trigeminal neuropathic pain.
