The majority of studies on pain processing have used, as a behavioral model of pain, the withdrawal reflex threshold of a limb provoked by a brief noxious stimulus, usually thermal, mechanical, or electrical. Although such traditional nociceptive tests have considerably improved our knowledge of pain mechanisms, the relevance of these models for understanding activity in chronic pain states is somehow limited. Indeed, pain associated with injury and disease is long lasting and frequently associated with inflammation and psychological distress. The molecular and cellular changes observed in nociceptive modulatory circuits provide evidence of this complex functional reorganization occurring in the presence of persistent pain [55, 56]. Furthermore, there is a clear indication that suppression of phasic and tonic pain can involve distinct mechanisms [57, 58]. In this regard, the formalin test was developed as a model of tonic pain that has a greater relevance to clinical pain than an acute phasic pain test [42, 46]. The local injection of formalin into the hindpaw produces a characteristic biphasic behavioral response [44, 59]. The first phase of activity reflects a direct activation of sensory afferents and may thus be of limited value in understanding mechanisms involved in persistent pain. A particular focus of the present study was therefore on the second phase, which corresponds to ongoing peripheral activity and involves central sensitization [42, 43].
We have recently implicated NTS1 receptors in NT's analgesic effects in tonic spinal pain paradigms . However, the experiments performed with the selective NTS1 antagonist SR48692 have suggested that another NT receptor may participate to the regulation of pain signaling, SR48692 partly reversing the analgesic effects of the non-discriminative agonist NT69L. To our knowledge, the present investigation demonstrates for the first time that NTS2 agonists given intrathecally produce antinociception in the formalin model of persistent inflammatory pain (Fig. 1). These results are in agreement with previous studies suggesting that NTS2 receptors play a role in the regulation of spinal nociceptive inputs. Thus, administration of the selective NTS2 agonists, JMV-431 and levocabastine, into the subarachnoid space of the lumbar spinal region elicited analgesia in the tail-flick test [15, 23]. The dose effect seen here with levocabastine was previous reported in different acute and tonic pain paradigms [15, 18]. The association of NTS2 receptors with different sub-populations of DRG neurons and their presence within the dorsal horn of the spinal cord also reinforced the involvement of NTS2 receptors in the control of incoming nociceptive messages. Indeed, NTS2 receptors are expressed by substance P- and lectin B4-containing small ganglion cells as well as by large-sized NF200-positive neurons documented to carry primary nociceptive sensory modalities and proprioceptive allodynia, respectively . A dense plexus of NTS2-immunoreactive neuronal processes and numerous NTS2-expressing nerve cell bodies were also detected in the superficial laminae of the dorsal horn and in deeper layers III-V, suggesting that NTS2 may play an important role in modulating the activity of spinal neurons .
Since the introduction of the formalin test, different approaches have been used to assess formalin-induced pain-related behaviors [42, 50]. The initial behavioral scoring in rats involved assigning of weighted scores to different kind of behaviors, the intensity of each nociceptive-related activity being directly correlated in proportion to their category weights [43, 44]. Based on the concept that the different behavioral endpoints represent a one-dimensional view of nociception, this method reflects the overall pain experience of the animal being tested. Using the weighted average pain scores, we found that NTS2-selective agonists and non-discriminative compounds gave rise to different action profiles during the inflammatory phase of the formalin test. Indeed, both JMV-431 and levocabastine had higher nociceptive scores than did NT69L (which binds to both NTS1 and NTS2) during the early phase 2 (21–39 min), while all drugs were efficient in reducing pain during the late phase 2 (40–60 min) (Fig. 2). Accordingly, we recently demonstrated that the selective-NTS1 agonist, PD149163 behaved as NT69L, suppressing pain-related behavioral activity in both early and late parts of phase 2 . The decrease in the analgesic efficacy of NTS2-selective agonists observed during the first half of phase 2 may be related to the non-involvement of NTS2 during the early tonic inflammatory responses. This hypothesis is supported by previous findings showing that some receptors involved in the regulation of pain processing differently modulate the early and late tonic phases 2 [45, 60]. For example, the synthetic opioid analog, loperamide, only abolishes flinching behaviors in the second part of phase 2 . Alternatively, it may also reflect the active participation of NTS2 receptors during the interphase. We could indeed speculate that the endogenous release of NT acting on NTS2 sites during the interphase may reduce or mask the analgesic effects of NTS2 agonists at the beginning of the inflammatory phase. This delay in the antinociceptive effects of NTS2-selective agonists possibly corresponds to the time frame required for functional resensitization of NTS2 receptors. This second hypothesis is reinforced by previous studies showing that the interphase of the formalin test is the result of active endogenous pain-inhibitory mechanisms [61–67].
Aside from the weighted-scores technique, pain intensity elicited by formalin injection may also be rated by the assessment of each individual nociceptive behavior, such as the total time the paw is kept elevated from the floor [68, 69], the number of times the rat flinches in a given time period [57, 59], or the time spent licking and biting the injected paw . A problem which may occur when measuring only one behavior such as lifting is that if a given treatment prevents this behavior, it is considered to produce 100% analgesia, despite the possibility that the rat may still exhibit other characteristic nociceptive behaviors, such as licking and biting . However, the advantage of single parameter recordings over the weighted-scores method is that by taking into account the measurement of multiple behaviors, it would be possible to assess nociceptive behaviors in a manner that reflects the multidimensional nature of pain experience [42, 46, 50, 59]. Based on this method of analysis, we demonstrated that NTS2-selective agonists behaved differently over the inflammatory phase than non-discriminative compounds and could therefore influence pain transmission in a distinctive manner. Thus, spinal delivery of JMV-431 and levocabastine significantly attenuated formalin-induced lifting but not licking, biting and shaking responses during the inflammatory phase (Fig. 3). Unlike NTS2 drugs, the NT69L agonist, acting on both NTS1 and NTS2 receptors, was able to reverse all nociceptive endpoint behaviors observed following tissue injury by intraplantar formalin (Fig. 4). In Sprague-Dawley rats, these stereotypic behavioral reactions to formalin have been used extensively to evaluate analgesic properties of drugs, and different effects on lifting, shaking, licking, and biting have also been observed with other agents [46, 50]. Systemic naloxone was indeed shown to increase formalin-induced flinching while simultaneously decreasing licking behavior . Furthermore, peripheral injection of NMDA receptor antagonists significantly attenuated formalin-induced lifting, while flinching behavior was not affected . This diversity of outcomes observed in the formalin test was also reported with other spinally administered analgesics and antidepressants [73, 74]. Even so, concerns regarding motor effects influencing formalin behaviors have been expressed. It has been advocated that licking/biting behaviors are subjected to motor influences and stereotypy, whereas flinching (e.g. paw elevations and paw shakes) is a more spontaneous response, being less influenced by other non-nociceptive behavioral changes (e.g. motor) [42, 46, 59]. In this study, we focused on doses that produced no observable motor impairment . Consequently, it is unlikely that the suppression of licking/biting behaviors induced by medullary delivery of non-selective NTS1/NTS2 compounds is related to alterations in motor performance.
The differential effects of pure NTS2 agonists in eliciting lifting and biting/licking responses noted here also suggest that theses behaviors involve distinct mechanisms or neural circuitry. Accordingly, pain-related behaviors can be associated with distinct brain structures, including spinal, brainstem, and cerebrally mediated responses to nociceptive stimulation . Well-known for evaluating the effects of analgesic treatments on long-lasting pain, the formalin test also allows the recording of behavioral reactions relayed at different levels of the central nervous system . Aside from the spinal cord, formalin-evoked responses seem to be mainly integrated in the brainstem and midbrain, as decerebration and decortication do not affect a range of behaviors [77, 78]. Regarding the stereotypical nocifensive behaviors observed following formalin administration, it was shown that the persistence of limb flexion is largely a spinal reflex, as it can be evoked following chronic spinalization . Conversely, supraspinal influences have been proposed to contribute in licking/biting behavioral manifestations associated with pain [74, 80, 81]. With this perspective, we could therefore hypothesize that NTS2 agonists, reducing spontaneous pain manifested by paw lifting, may specifically recruit spinal endogenous antinociceptive systems. Alternatively, the cumulative effects of NT69L (which binds to both NTS1 and NTS2) on lifting and licking/biting behaviors may reveal that NTS1 receptor activation modulates supraspinal nociceptive networks following formalin-induced tissue injury.
Persistent nociceptive stimulation of primary sensory afferents causes prolonged alterations in the neurochemistry and phenotype of postsynaptic dorsal horn neurons. It enhances their excitability leading to hyperalgesia and allodynia that develop after tissue injury. Accordingly, noxious stimuli initiate specific spatio-temporal patterns of c-fos expression within the spinal cord, allowing mapping of spinal nociresponsive neuronal populations (including projection cells and both excitatory and inhibitory interneurons) [53, 54]. In the present study, we used c-fos expression as an anatomical correlate of the behavioral measures to evaluate the antinociceptive activity of NTS2 compounds. Our results revealed that intrathecal NTS2 agonists reverse formalin-induced activation of c-fos-like immunoreactivity within the dorsal horn (Figs. 5, 6). Specifically, we found that JMV-431 and levocabastine significantly reduced c-fos-LI within the superficial layers of the rat spinal cord (Rexed's laminae I-II), and also exerted inhibitory effects on noxiously evoked c-fos expression in deeper laminae. In view of its spinal distribution , NTS2 receptor-mediated inhibition of nociceptive transmission may occur at multiple laminae levels.
Converging evidence supports the idea that noxious inputs to the spinal cord are actively modulated by descending systems that serve not only to inhibit, but also to facilitate nociception . The participation of these suprapinal sites is of great clinical significance as it is now apparent that the development and maintenance of exaggerated pain states may result from an alteration of the balance between facilitatory and inhibitory brain circuits . During the formalin response, the descending pain-control influences originating from the rostroventromedial medulla are indeed essential in maintaining the nociceptive behavioral manifestations . The mechanisms by which NT receptor agonists induce analgesia may therefore arise from local spinal pain inhibition or regulation of supraspinal inputs. As spinal activation of NTS2 receptors was shown here to block specifically the nociceptive behaviors integrated at the spinal cord level, we might propose that spinal delivery of NTS2 agonists interrupts conduction of the nociceptive processing by reducing excitability of peripheral primary afferent terminals and inhibits dorsal horn c-fos nociresponsive interneuron activity. However, since non-discriminative agonists reversed supraspinal nocifensive behaviors, the antinociception observed could also rely on the inhibition of the spino-bulbo-spinal loop by NTS1 receptor activation. It is currently thought that this inhibitory drive is due to a reduction in the activity levels of c-fos-positive neurons projecting to upper brainstem structures .
To our knowledge, the present study is the first to demonstrate that activation of NTS2 receptors induces analgesia in a persistent pain model. The intrathecal administration of NTS2 drugs was shown to reduce behavioral and molecular markers of persistent inflammatory pain. Our results also suggest that a dichotomy exists between the spinal antinociceptive actions of NTS2-selective agonists and NTS1/NTS2 mixed compounds. Further research to design NTS1 and NTS2-selective novel analogs that cross the blood-brain barrier may therefore offer new avenues for the treatment of chronic pain.