The main results of this study point out that AEA (0.5–50 pmol), a TRPV1/CB1 agonist, injected into the mouse dPAG does not produce any intrinsic effect on acute pain as assessed through the tail-flick test (Exp. 1A). Conversely, intra-dPAG injections of capsaicin (0.1 and 1 nmol – Exp. 1B) or WIN (10 and 50 nmol – Exp. 1C) provoked a marked TRPV1- and CB1-dependent antinociception, respectively. While the blockade per se of CB1 receptors did not change nociceptive response (Exp. 1D), intra-dPAG injection of 6-IODO, a TRPV1 antagonist, produced a hypernociceptive effect (Exp. 1E). Interestingly, under blockade of CB1 receptors, intra-dPAG AEA produced a clear and consistent antinociceptive effect (Exp. 2). In contrast, the blockade of TRPV1 did not change the lack of effects of intra-dPAG injection of AEA on nociceptive response (Exp. 3).

Intra-dPAG injections of AEA (0.5–50 pmol), an endocannabinoid/endovanilloid agonist (Zygmunt et al., 1999; Van Der Stelt and Di Marzo, 2004; Marinelli et al., 2007), failed to alter TFL throughout the 40-min test. This unexpected result contrasts with previous findings showing that intra-vlPAG injections of mild doses of URB597 (an FAAH enzyme inhibitor), which in turn increases endogenous AEA, led to a vanilloid-mediated analgesia in rats exposed to the plantar test (Maione et al., 2006). However, those authors also reported that local low or high doses of the FAAH inhibitor provoked CB-mediated hyperalgesia (Maione et al., 2006). Based on present study, AEA acting specifically on CB substrates (Exp. 3) corroborated the lack of effect of AEA (Exp. 1A) suggesting that the doses of this vanilloid/CB agonist used in Exp. 1A might also have stimulated preferentially CB1 receptors. Indeed, the role of CB substrates on nociception is under a complex debate. A less likely possibility might account for this lack of effect considering a combined stimulation of TRPV1/CB1 receptors in Exp. 1, since it has been demonstrated a net null effect following both vanilloid/CB activation on the ventrolateral column of the PAG (Maione et al., 2006). In this regard, since previous results from our laboratory showed a vanilloid-mediated antinociception into the dorsal PAG (Mascarenhas et al., 2015), present study was conducted accordingly in the same midbrain column in an attempt to be consistent with previous findings. These site differences (dorsal versus ventrolateral columns) could also explain such discrepancies.

To clarify the involvement of dPAG TRPV1 and CB1 receptors in the modulation of the nociceptive response this study investigated the effects of specific vanilloid or CB agonists into the mouse dPAG on nociception. Interestingly, both capsaicin (0.1 and 1 nmol) and WIN (10 and 50 nmol) injected into the mouse dPAG increased dose-dependently the TFL of mice, suggesting a vanilloid- and a CB-dependent antinociception, respectively. Present study also reported a delayed CB-mediated antinociception following a local injection of a low dose (1 nmol) of WIN. In general, these results corroborate many other studies showing that capsaicin injected into various PAG columns provokes antinociception in rodents (Palazzo et al., 2002; Starowicz et al., 2007; Mascarenhas et al., 2015) possibly by facilitating the descending inhibitory system. Regarding CB1 role in the modulation of nociception, as presently shown with intra-dPAG injections of WIN, a great body of evidence have shown that phytocannabinoids (Lichtman and Martin, 1996; Meng et al., 1998), synthetic agonists (Meng et al., 1998), and endocannabinoids (Maione et al., 2006) induce antinociception as assessed through acute pain tests. However, contrasting results have also been reported with TRPV1 and CB1 manipulations. For instance, the hypernociception reported following vanilloid stimulation (McGaraughty et al., 2003) is an effect that has been attributed to the capacity of desensitization of TRPV1 receptors (Palazzo et al., 2008). In this later scenario, i.e., vanilloid-induced hypernociception, only high doses of TRPV1 agonist would be sufficient to induce desensitization (Palazzo et al., 2008). Regarding the hypernociceptive effects following CB1 receptor activation, previous studies have demonstrated immediate and delayed hypernocicetive effects when low doses of WIN or URB597 were injected into the rat vlPAG (Maione et al., 2006). This paradoxical profile of CB1 agonists seems to be due to the expression of CB1 receptors on both glutamatergic and GABAergic neurons at pain modulating circuitry of these murine species (Palazzo et al., 2008).

Accordingly, Exps. 1D and E were carried out in order to find intrinsically inactive doses of CB1 and TRPV1 antagonists on nociception when injected into the mouse dPAG. Exp. 1D revealed that AM251, a CB1 receptor antagonist, did not change nociceptive response of mice subjected to the tail-flick test, indicating lack of CB tonic control over dPAG CB1 receptors in the modulation of acute pain. This is in accordance to the fact that the CB system is recruited only on demand [for a review see Morena and Campolongo (2014) and Ulugol (2014)]. On the contrary, Exp. 1E showed that intra-dPAG 6-IODO (TRPV1 antagonist) at the highest dose (3 nmol) decreased the TFL, suggesting a hypernociceptive effect and therefore an endovanilloid tonus within the mouse dPAG modulating nociception. Starowicz et al. (2007) have first demonstrated a tonic endovanilloid facilitation of glutamate release within rats PAG, since 5′-iodoresiniferatoxin, a selective TRPV1 antagonist, facilitated nociceptive responses. In addition, authors have demonstrated that endovanilloids contribute to anxiety modulation. In this context, capsazepine (TRPV1 antagonist) injected into the PAG attenuated the defensive behavior of rats exposed to the elevated plus maze (Moreira et al., 2007).

The well-known CB1-mediated antinociception found in Exp. 1C seems to be due to the inhibition of GABA release from PAG interneurons, which in turn would contribute to disinhibition of PAG antinociceptive outputs (Moreau and Fields, 1986; Meng et al., 1998; Vaughan et al., 2000). It contrasts with the lack of effect of AEA acting specifically on CB substrates reported on Exps. 1A and 3. In this context, although no Experiment has assessed the said issue, it is likely that the doses of AEA stimulated both neurons population-expressing CB1 receptors and nociception might have been physiologically counterbalanced which accounts for the lack of effect. In addition, WIN displays different pharmacodynamics aspects (higher CB1 affinity therefore lower Ki value) and it is not subjected to FAAH hydrolyses compared to AEA, which makes difficult the comparison of doses of a synthetic versus endogenous CB1 agonist. Furthermore, it is still necessary considerably more investigation to unravel the participation of glutamatergic and GABAergic neurons-expressing CB1 receptors since, physiologically, CB-mediated outcomes are under opposite pathways.

Attempting to determine the role of exogenous AEA binding preferentially at either TRPV1 or CB1 receptors, Exps. 2 and 3 consisted on evaluating nociceptive response in mice that had received intra-dPAG injection of AM251 (10 pmol) or 6-IODO (1 nmol) prior to local injections of AEA. Interestingly, contrasting with the results obtained in Exp. 1A, AEA (5.0 and 50 pmol) produced a marked antinociceptive effect only under the blockade of CB1 receptors (Exp. 2). AEA, at the highest dose, provoked a high-magnitude 20-min lasting antinociceptive effect possibly via vanilloid substrates, corroborating Exp. 1B, which revealed a similar capsaicin-induced antinociception. On the contrary, the blockade of vanilloid substrates failed to reveal any intrinsic effect of intra-dPAG AEA on nociception (Exp. 3), corroborating the lack of effect observed in Exp. 1A, where only AEA was injected. Thus, it seems reasonable to suggest that the doses of AEA used in Exp. 1A have also preferentially stimulated the CB substrates even though no antagonism was performed. In this context, in terms of pharmacodynamics, AEA binds to CB1 receptors with higher affinity (Ki value between 37 and 116 nM) compared to TRPV1 receptor (Ki value 1.66 μM) (Ross et al., 2001). Therefore, Experiment 1A seems to corroborate the higher affinity of AEA to CB substrates at a behavioral level since it revealed the same outcome of Exp. 3.

In this context, it has been shown that AEA might also act via other ligand-gated channels, such as the 5-HT3 and glycine receptors, which potentially could contribute to AEA-induced effects on nociception. However, the AEA role on 5-HT3 receptors highly depends on the abundance of this receptor at the cell surface of specific brain sites (Barann et al., 2002; Xiong et al., 2008). Furthermore, AEA seems to enhance glycine clearance in the synaptic cleft (Pearlman et al., 2003), which, physiologically, could impair the NMDA-dependent excitatory current and by extension AEA-induced antinociception which depends on glutamate release into RVM OFF cells. These evidence have weakened the concern with the potential influence of other channels in AEA effects together to the fact that several evidence point out the relevance of vanilloid/CB substrates in mediating AEA-induced effects (Zygmunt et al., 1999; Vaughan et al., 2000; Fenwick et al., 2017). In other words, these evidence strengthen the fact that Exps. 2 and 3 indeed led AEA to bind with specificity to vanilloid and CB substrates, respectively.

According to our hypothesis, AEA-induced antinociception seems to be easily reproduced when it is vectored to vanilloid substrates where only desensitization must be avoided. In other words, controlling the amount of endovanilloid that stimulates TRPV1 receptors, AEA will ultimate provoke antinociception. In contrast, aiming at studying CB1-mediated antinociception, one must take into account the complex neurophysiology of the CB substrates within the dPAG, where both glutamatergic and GABAergic neurons express CB1 receptors. In addition, it seems that while present results unmasks exogenous AEA antinociceptive effects after CB1 blockade in the PAG, FAAH inhibitors unravel endogenous AEA effects intra-vlPAG that might be pro- or antinociceptive depending on the doses (Maione et al., 2006).

Noteworthy, in Exp.1 the dosage of AEA (0.5–50 pmol) was a bit disconnected to that used for capsaicin (0.01–1 nmol). On the one hand, it might weaken our study in view of their different affinities for TRPV1 (AEA displays lower affinity to TRPV1 than capsaicin) and therefore a higher dosage of AEA and a lower dosage of capsaicin than those employed in the present study should have been considered. However, on the other hand, it strengthens the fact that a very low dose of AEA was able to induce antinociception via TRPV1 under local CB1 blockade (Exp. 2), corroborating the hypothesis that depending on the substrates (i.e., TRPV1 or CB1) AEA might provoke a clear and potent antinociceptive effect.

Finally, it is likely that a broader range of AEA doses on Exp. 3 would possibly reveal a different outcome of this compound acting preferentially on CB substrates, since the doses of AEA determine the proper CB1-mediated effects. However, analyzing the effects in Exp. 2, present data suggest an interesting approach to address exogenous AEA effects on nociception (i.e., vectoring AEA to act preferentially on vanilloid substrates) and eventually allowing promising clinical trials which until now does not seem to translate to humans the potential of CB compounds (e.g., AEA) in pain management.

As far as we know, present results are the first to demonstrate antinociceptive effects of exogenous AEA injected into the dPAG specifically in an acute pain test, i.e., tail-flick test. So far, only endogenous AEA has been investigated in the modulation of nociception. In conclusion, present study revealed an antinociceptive effect of exogenous AEA injected into the dPAG only when CB1 receptors were antagonized, suggesting an important role of AEA in the vanilloid substrates that modulate acute pain within this midbrain area of mice. Therefore, it makes relevant to investigate further approaches considering the role of AEA binding specifically in vanilloid substrates as a potential new methodology to address acute pain on basic research and perhaps clinical trials.