This study presents behavioral evidence for opposing effects of endocannabinoids in responses to non-nociceptive mechanical vs. nociceptive chemical stimuli. Specifically, 2-AG and anandamide both enhanced responses to non-nociceptive stimuli and reduced responses to nociceptive stimuli. Both of these effects were blocked when the TRPV channel inhibitor SB366791 was co-injected. We also examined the effects of endocannabinoids on animals that had undergone an injury to the posterior sucker that produced persistent (at least 7 days) sensitization to both nociceptive and non-nociceptive stimuli. 2-AG and anandamide had no effect on responses to non-nociceptive stimuli in these injury-induced sensitized animals. This lack of effect is likely due to the fact that further decreases in the response threshold cannot occur because these animals already maximally sensitized due to the injury. However, it is also possible that endocannabinoids mediated the injury-induced sensitization itself and therefore this sensitization occludes any additional effects of injected 2-AG or anandamide. High frequency stimulation of the Hirudo nociceptive neurons does produce endocannabinoid-mediated potentiation of the pressure cell synapses that are likely to mediate responses to the von Frey fibers used in this experiment³³. Future studies will examine the potential role of endocannabinoid signaling in mediating injury-induced sensitization to non-nociceptive injury.

2-AG and anandamide injections increased the response latency to chemical nociceptive stimuli (AITC), an anti-nociceptive effect. Injury-induced sensitization was not observed following injection of 100 µM 2-AG and there was an increased response latency relative to non-injured control animals for several days after injection. 50 µM 2-AG produced a smaller anti-nociceptive effect, but one that still lasted several days. 25 µM 2-AG was ineffective. By comparison anandamide was less effective with only the highest concentration (100 µM) producing an anti-nociceptive effect that only lasted one day. Both 2-AG and anandamide produce similar levels of depression in Hirudo nociceptive synapses although this was only measured for 1–2 hrs^{31, 33, 35}. It is not clear at this time why 2-AG and anandamide are so different in terms of the duration of their behavioral effects.

These results are consistent with previous in vitro neurophysiological studies carried out in isolated ganglia and semi-intact preparation in which endocannabinoids depress nociceptive (N cell) synapses, but potentiate non-nociceptive (P cell) synapses (see Fig. 1A)^31–33. Pharmacological inhibitors of TRPV1 blocked both endocannabinoid-mediated potentiation of P cell synapses and depression of the N cell synapses. Previous pharmacological studies have shown that Hirudo possesses a TRPV-like channel both peripherally and in the CNS that responds to capsaicin and TRPV1 antagonists such as SB366791 and capsazepine^{24, 31, 33}. Depression of N cell synapses is due to activation of presynaptic TRPV-like channels that is thought to lead to a decrease in neurotransmitter release (see Fig. 1A) and is calcineurin- and transcription-dependent^{30, 34, 35}. These mechanisms are similar to endocannabinoid/TRPV1-mediated depression in hippocampal synapses^{41, 42}. Although also TRPV-mediated, P cells lack TRPV-like channels and potentiation of P cell synapses is an indirect process that involves disinhibition of theses synapses (see Fig. 1A)^{32, 33}. Nociceptive synapses are “protected” from this disinhibition because they are depolarized by GABA due to elevated levels of intracellular Cl^{− 33, 36}. At this time the GABAergic neurons in Hirudo that undergo this suggested endocannabinoid-mediated depression have not been identified.

An important caveat of these studies and much of the previous Hirudo work is that they are based on pharmacological manipulations. It is possible that the observed drug effects are due, at least in part, to off-target effects unique to invertebrates. For example, the platyhelminth Schistosoma mansoni, responds to capsaicin, but this effect is mediated by TRPA1 channels (which are also present in Hirudo)^{25, 40}. We have tried to minimize this potential confound in our past studies by utilizing multiple pharmacological agents^{30, 31, 33}. However, this concern will remain until Hirudo versions of these proteins are isolated and directly examined.

The current experiments, combined with previous synaptic studies, have relevance to understanding endocannabinoid/TRPV-based modulation of nociception. Endocannabinoids have been reported to exert an anti-nociceptive effect due at least in part to depression of glutamatergic transmission at primary afferent synapses in the spinal cord^{4, 5, 7, 43, 44}. However, stimulation of CB1 receptor can also enhance nociception due to depression of GABAergic/glycinergic inhibitory transmission within the spinal cord⁸. Interestingly, injury-induced allodynia due to TRPV1-mediated disinhibition in the spinal cord has been observed although it is not known whether endocannabinoids are activating the TRPV1 channel⁴⁵. Finally, pro-nociceptive effects are also observed in animals with a genetic knock-out of fatty acid amide hydrolase (FAAH), which is responsible for anandamide metabolism, and these effects were mediated by both CB1 and TRPV1⁹. This capacity for endocannabinoids to have both pro- and anti-nociceptive effects, potentially through both CB1- and TRPV1-mediated signaling, may help to explain why some clinical studies of cannabinoid-based therapies to treat chronic pain can sometimes result in either no effect or a worsening of symptoms^10–12. These pro-nociceptive effects may be due in part to endocannabinoids disinhibiting (and therefore enhancing) the nociceptive circuitry. Disinhibition is a critical mechanism that “opens the gate” for non-nociceptive afferents to have access to spinal nociceptive circuits^{15–17, 45}, but the mechanisms by which injury elicits disinhibition associated with sensitization of non-nociceptive pathways are not fully understood. Differences in Cl⁻ gradients between nociceptive and non-nociceptive afferents may contribute to the latter being more sensitive to disinhibition^{10, 33, 36, 46–48}. In Hirudo this disinhibition selectively affects non-nociceptive pathways, but may have effects on both non-nociceptive and nociceptive pathways in mammals possibly due to the more complex circuitry (in terms of multiple classes of inhibitory and excitatory interneurons) at the spinal cord level^{14, 15}.

Together, these findings demonstrate direct behavioral evidence for the opposing effects of endocannabinoids in both injured and non-injured animals. These results significantly contribute to understanding the potential role of endocannabinoids in both the induction and attenuation of pain conditions and demonstrate the need for more studies characterizing the specific mechanisms unique to sensitization of nociceptive sensory pathways versus sensitization of non-nociceptive pathways. From a clinical standpoint, endocannabinoid-based therapies may only have efficacy for certain types of pain conditions. Specifically, it is possible that cannabinoid-based therapies will be more effective for conditions that involve hyperalgesia, but either less effective for or perhaps even exacerbate conditions that include allodynia. Such considerations must be taken into account when designing future clinical studies that seek to use the endocannabinoid signaling system to treat chronic pain.