The main finding of this study is the anti-allodynic effect of SA, the main compound of Salvia Divinorum, in a formalin-induced chronic pain model in mouse. We have here coupled behavioural pharmacology with in vivo electrophysiology in order to: 1) clarify the long-lasting behavioural responses and NS neuronal activity induced by a single formalin injection 2) investigate the contribute of KOR/CB1 receptor manipulation in formalin-induced pain condition. In particular, we have shown that SA repeated treatment, reduces mechanical allodynia up to 7 days after formalin peripheral administration in mice. The local injection of formalin into the hind-paw of rodents has been considered mainly a model of persistent inflammatory pain, which induces a characteristic nocifensive biphasic response, oedema and inflammation . We have previously demonstrated that a single administration of SA partially reduced the first phase of formalin-induced nociceptive behaviour, which is caused by direct activation of nociceptive sensory afferents, and it completely abolished the second one  which is due to the release of inflammatory mediators and it is also associated with central sensitization [33–35]. However, peripheral formalin injection also exerts long-term nerve injury [15, 17, 36, 37]. Accordingly, we have shown that formalin leads to tactile allodynia ipsilaterally to the injection side spreading also to the contralateral hind-paw. SA was effective in reducing the formalin-induced ipsi- and contralateral allodynia up to 7 day treatment. Moreover, we found that SA effects were prevented by KOR and CB1R selective antagonists. Indeed, the involvement of CB1 receptor in SA-mediated effects has been already shown in other inflammatory animal models [30, 38] and also in emotional behaviour studies in rodents . How SA can reduce pain transmission by modulating CB receptors is presently unclear. SA shows very low or no affinity for CB1 and CB2 receptors, respectively, and it seems to be unable to modulate significantly endocannabinoid levels [38, 40]. The finding that the effect of U-50488, the synthetic KOR agonist, was counteracted by nor-BNI and AM251 similarly to SA, suggests a possible functional KOR-CB1R interaction rather than an exclusive effect of the drug. Several findings have suggested that a functional cross-talk between cannabinoid and opioid systems exists [41–43], even if the mechanisms of such interaction are not still clear. A direct receptor–receptor interaction, such as heteromerization, as already suggested by Rios et al.,  for CB1R and MOR opioid receptor, could represent one of possible mechanisms. Indeed, cannabinoids and opioids mutual interaction has been described with particular emphasis on the pivotal role of opioid receptors and peptides in cannabinoid-mediated analgesia [41, 43] and importantly, such a functional interaction between the two systems seem to be altered during chronic pain . In our model, we found a significantly increase of spinal CB1 receptor expression, while no changes have been detected for CB2R or KOR. The involvement of the CB2R, which is known to play an important role in chronic inflammatory process and also in reducing pain behaviour associated with neuropathic pain [13, 46–48] has been also object of this study. The co-administration of SA with AM630, the selective CB2 receptor antagonist, ruled out the contribute of this receptor in SA-mediated effects.
Inflammation, tissue damage or nervous system injury result in the activation of immune and inflammatory cells, leading to consequent production of several mediators able to increase pain hypersensitivity. By using immunohistochemical approaches, we have found that formalin induces an increase of Iba-1 staining, a marker of microglia cells, in the dorsal horn of lumbar spinal cord ipsilaterally to the injection, as previously demonstrated [15, 49]. In this study we also have observed an increase of hypertrophic GFAP positive cells 7 days post-injection in the same area. Astrocytes and microglia cells showed a typical morphological shape corresponding to the activated phenotype. The effectiveness of SA in alleviating mechanical allodynia matched with a reduction of the activated cells number and with the modulation of IL-10 and iNOS, two different markers involved in the inflammatory processes. In this study we found that, at time point evaluated, the IL-10 protein level was increased by SA treatment in terms of intracellular production compared to saline or formalin treated animals, as revealed by immunohistochemestry, while it was normalized in the released form, as revealed by ELISA. These data suggest that the mechanisms leading to the long-term nocifensive behaviour induced by formalin are correlated with a reorganization of the spinal cellular populations in which microglia and astrocytes activation play a significant role. In this context, the effectiveness of SA in reducing the activation of astrocytes and even more of microglia seems to be mainly mediated by CB1R rather than KOR. Indeed, we found that chronic treatment with AM251, but not nor-BNI, (see Additional file 1: Figure S1 and Additional file 2: Figure S2) reverted the effect of SA in reducing microglial activation. These data are consistent with the evidence showing that cannabinoids are more effective than opioids in alleviating neuropathic pain .
Importantly - and consistently with the behavioural data - electrophysiological experiments revealed that SA caused a strongly reduction of evoked activity of NS neurons in formalin-injected mice, suggesting that KOR/CB1R activation plays an important role in the transmission of noxious signals to the spinal cord under pathological conditions. Indeed, our findings are the first one showing the long-lasting effects of formalin injection on NS neurons activity. The decreased threshold of activation and the increased responsiveness to mechanical noxious stimuli of NS neurons found 7 days after formalin, suggest that a single peripheral formalin injection, beyond motor dysfunctions, induces a central sensitisation, similarly to a neuropathic pain condition induced by nerve injury. The finding that 7 days are needed to modulate spinal neuronal activity seems to be in contrast with allodynia development and with the lack of changes in spinal pro-inflammatory mediator level, observed already 3 days post-injection. However, this may be due to the formalin-induced intense inflammation with an immediate production of prostaglandins and accumulation of neutrophils and infiltrating mononuclear cells which could be alone responsible for the induction of allodynia. Indeed, a range of peripheral inflammatory stimuli, such as the peripheral injection of carageenan, mustard oil and complete Freund’s adjuvant, have been shown to induce a robust allodynia and hyperalgesia in adult animals [51, 52]. In our model, the initial microglial activation and/or the release of pro-inflammatory cytokines appears to be unable to sensitize dorsal horn sensory neurons. It would be reasonable to hypothesize that microglia and astrocytes can participate together in neuronal sensitization, not only by releasing cytokines and various mediators, but also more directly via release of glutamate and/or by evoked-changes in synaptic ion currents . The evidence of the crucial role of astrocytes in the maintenance of mechanical allodynia in chronic pain and our finding that astroglial activation does not occur until to the 7th day after formalin injection, supports their involvement in the establishment of central sensitization. Moreover, some brain areas of the endogenous antinociceptive pathway, i.e. periaqueductal gray (PAG)-rostral ventromedial medulla (RVM) axis, could be activated in pathophysiological conditions associated with persistent pain, so reducing neuronal sensitization at spinal level. In fact, previous findings showing an increased endocannabinoid release within the dorsal and lateral PAG following a single formalin injection , suggest a possible activation of CB1 receptors in such areas in inflammatory conditions, which could be responsible for the inhibition of the spinal NS neurons over-excitability.